1 Introduction

2 Materials and methods

2.2

2.2 Study population

The study population included patients aged 18–84 years who underwent ACDF, defined by the following CPT and ICD codes: CPT-22551, 22554, 22552, 22585; ICD-9-P-8102; and ICD-10-P codes ranging from 0RG1070 to 0RG24Z0. Subsequent tracheostomy was identified using CPT-31600, 31603; ICD-9-P-311; and ICD-10-P codes ranging from 0B110F4 to 0B114Z4, covering the years 2010 to 2022Q2. Patients were then stratified into 3 groups based on the timing of tracheostomy after ACDF: early (0–5 days), intermediate (6–10 days), and late (11–20 days), as shown in Fig. 1. The time windows of this study were selected based on reported average or median time lengths from ACDF to tracheostomy placement from available prior studies, which mostly fell within the 15-day time window.^10–14 Patients with a history of infection or malignancy of the cervical spine were excluded from the population. Any posterior approach was excluded. Baseline comorbidities were assessed with the Charlson Comorbidity Index (CCI) and include obesity (BMI 30–40 and BMI 40+), smoking, diabetes mellitus, chronic kidney disease, depression, osteoporosis, alcohol use, chronic obstructive pulmonary disease (COPD), bronchiectasis, asthma, dementia, Parkinson's disease, congestive heart failure (CHF), and cancer. Additionally, for surgical parameters, the patient population studied was evaluated on whether they experienced spinal cord injury (SCI) and the level of cervical fracture (C3-C5, C6-T1). Patients were also assessed on whether they received a corpectomy or had intracranial hemorrhage, traumatic brain injury (TBI), vertebral artery dissection, acute respiratory failure, and ventilator use 14 days prior to tracheostomy. A flow chart of the study design is shown in Fig. 2.

Close

Fig. 1 Timeline of Tracheostomy Timing. ACDF, anterior cervical discectomy and fusion.

Export to PPT

Close

Fig. 2 Tracheostomy Timing and ACDF Study Design Flowchart. ACDF, anterior cervical discectomy and fusion.

Export to PPT

3 Results

A total of 594 patients with sub-axial cervical fractures who underwent anterior cervical discectomy and fusion (ACDF) were included and stratified by tracheostomy timing into early (0–5 days; n = 147), intermediate (6–10 days; n = 271), and late (11–20 days; n = 176) cohorts.

3.3

3.3 Postoperative complications

As shown by Table 2 and Fig. 3, intermediate tracheostomy timing was associated with significantly lower odds of pneumonia compared to early tracheostomy (54.2 % vs. 63.3 %; OR 0.59, 95 % CI 0.38–0.93; P = 0.022). Additionally, late tracheostomy was associated with significantly lower odds of CVA than the early group (6.1 % vs. 1.1 %; OR 0.05, 95 % CI 0.00–0.40; P = 0.014). No other medical complications, including sepsis, urinary tract infection, or PE, were significantly associated with tracheostomy timing. As shown in Fig. 4, surgical complication rates did not significantly differ among groups. No significant differences were found in rates of wound infection, dysphagia, recurrent laryngeal nerve palsy, or ventilator dependence.

Table 2 Multivariate analysis of 90-day postoperative complications.

	Early (0-5d) n = 147		Intermediate (6-10d) n = 271		Late (11-20d) n = 176		Intermediate vs Early OR (95 % CI)	P value	Late vs Early OR (95 % CI)	P value
Medical Complications
Myocardial infarction	2	1.4 %	6	2.2 %	4	2.3 %	4.18 (0.51–70.61)	0.235	2.82 (0.29–46.02)	0.398
Cardiac arrest	10	6.8 %	20	7.4 %	15	8.5 %	1.12 (0.48–2.72)	0.794	1.31 (0.52–3.41)	0.566
Atelectasis	80	54.4 %	131	48.3 %	89	50.6 %	0.64 (0.41–1.00)	0.051	0.69 (0.42–1.13)	0.145
Pleural effusion	60	40.8 %	104	38.4 %	76	43.2 %	0.87 (0.56–1.36)	0.564	0.94 (0.57–1.54)	0.811
PE	4	2.7 %	7	2.6 %	7	4.0 %	0.65 (0.13–3.42)	0.591	0.66 (0.14–3.25)	0.598
DVT	10	6.8 %	27	10.0 %	20	11.4 %	1.43 (0.65–3.40)	0.385	1.48 (0.61–3.74)	0.39
CVA	9	6.1 %	9	3.3 %	2	1.1 %	0.51 (0.15–1.69)	0.273	0.05 (0.00–0.40)	0.014∗
Seizure	2	1.4 %	1	0.4 %	0	0.0 %	–	–	–	–
Ileus	14	9.5 %	22	8.1 %	20	11.4 %	0.75 (0.35–1.62)	0.459	1.21 (0.53–2.81)	0.641
Electrolyte abnormality	33	22.4 %	65	24.0 %	47	26.7 %	1.08 (0.64–1.82)	0.767	1.09 (0.61–1.94)	0.766
Acute kidney injury	23	15.6 %	41	15.1 %	26	14.8 %	0.94 (0.50–1.79)	0.848	0.64 (0.30–1.34)	0.240
Pneumonia	93	63.3 %	147	54.2 %	104	59.1 %	0.59 (0.38–0.93)	0.022∗	0.80 (0.48–1.33)	0.396
Sepsis	25	17.0 %	39	14.4 %	23	13.1 %	0.74 (0.41–1.36)	0.337	0.54 (0.27–1.08)	0.085
UTI	62	42.2 %	101	37.3 %	57	32.4 %	0.79 (0.51–1.24)	0.315	0.65 (0.43–1.09)	0.109
Surgical complications
Transfusion	10	6.8 %	10	3.7 %	8	4.5 %	0.40 (0.14–1.18)	0.097	0.49 (0.14–1.59)	0.244
Dysphagia	81	55.1 %	135	49.8 %	78	44.3 %	0.87 (0.56–1.35)	0.558	0.66 (0.40–1.09)	0.112
RLN palsy	2	1.4 %	6	2.2 %	2	1.1 %	2.30 (0.20–40.95)	0.523	1.28 (0.07–25.54)	0.859
Wound dehiscence	0	0.0 %	9	3.3 %	5	2.8 %	–	–	–	–
Hematoma	1	0.7 %	0	0.0 %	1	0.6 %	–	–	–	–
Deep wound infection	3	2.0 %	9	3.3 %	8	4.5 %	1.67 (0.42–8.58)	0.491	2.04 (0.46–11.32)	0.367
Superficial wound infection	1	0.7 %	6	2.2 %	3	1.7 %	3.61 (0.38–99.06)	0.328	3.02 (0.27–82.96)	0.408
Anterior reoperation	1	0.7 %	1	0.4 %	2	1.1 %	–	0.813	–	0.426
Ventilator dependence	55	37.4 %	98	36.2 %	65	36.9 %	0.80 (0.49–1.32)	0.401	0.61 (0.35–1.08)	0.092
Secondary Outcomes
ED visit	26	17.7 %	52	19.2 %	26	14.8 %	1.06 (0.61–1.86)	0.826	0.84 (0.43–1.62)	0.616
Hospital readmission	74	50.3 %	126	46.5 %	79	44.9 %	0.81 (0.52–1.25)	0.348	0.77 (0.47–1.26)	0.314
Mortality	25	17.0 %	35	12.9 %	31	17.6 %	0.65 (0.32–1.33)	0.243	0.61 (0.28–1.34)	0.224

Close

Fig. 3 90-day Medical Complications Forest Plot. A, Intermediate versus early tracheostomy group. B, Late versus early tracheostomy group. Odds ratios and 95 % confidence intervals were plotted on a logarithmic scale. PE, pulmonary embolism; DVT, deep vein thrombosis; CVA, cerebrovascular accident; UTI, urinary tract infection.

Export to PPT

Close

Fig. 4 90-day Surgical Complications Forest Plot. A, Intermediate versus early tracheostomy group. B, Late versus early tracheostomy group. Odds ratios and 95 % confidence intervals were plotted on a logarithmic scale. RLN, recurrent laryngeal nerve.

Export to PPT

3.4

3.4 Health utilization and mortality

As shown by Table 2 and Fig. 5, no statistically significant associations were observed between tracheostomy timing and 90-day hospital readmission, emergency department visits, or mortality.

Close

Fig. 5 90-day Secondary Outcomes Forest Plot. A, Intermediate versus early tracheostomy group. B, Late versus early tracheostomy group. Odds ratios and 95 % confidence intervals were plotted on a logarithmic scale. ED, emergency department.

Export to PPT

4 Discussion

Tracheostomy is often applied following anterior cervical spine surgery to support ventilation and postoperative care. Currently, limited research endorses early tracheostomy as a safe procedure following surgery. This research employed a large database to study the 90-day medical and surgical complications following ACDF with varying postoperatively tracheostomy placement windows. The tracheostomy time frames were chosen based on prior studies to capture the identified average time frames to post-fixation tracheostomy, ranging from the day of ACDF to approximately 11 days post-ACDF.^10–14

Regarding patient demographics, significantly different rates of COPD and CHF at baseline were observed among the three groups. Patients in the late tracheostomy group were more likely to have COPD and CHF when compared to the other groups. Additionally, the late group was found to have a higher usage of ventilators at 52.3 % in the two weeks leading up to their tracheostomy. COPD and CHF may contribute to a longer ventilator use status for trauma patients. A prior study has shown that the majority of patients (59.3 %) requiring prolonged mechanical ventilation for more than 2 weeks in the respiratory care unit had COPD.¹⁵ In another study, COPD patients were found with a higher risk of prolonged mechanical ventilation of more than 21 days than their non-COPD control group.¹⁶ In the same study, CHF was also identified as one of the significant co-comorbidities that significantly increased the risk for prolonged mechanical ventilation. Subsequently, given anticipated prolonged ventilator use, it is possible that the care-providing team decide to transition patients from endotracheal intubation to tracheostomy.¹⁷ It is possible that in the late group, given higher rates of baseline comorbidities such as COPD and CHF, patients require further escalation of airway management to need mechanical ventilation although no significant differences were seen in acute respiratory failure.

In comparison, the early tracheostomy group was found with a higher rate of vertebral artery dissection at 17 % prior to tracheostomy, with significant differences among the three groups. The correlation between vertebral artery dissection and cervical trauma has been established with variable incidences, ranging from 0.49 % in the general blunt trauma cases to 46 % in mid-cervical spine fracture or subluxation patients.^18,19 The presence of vertebral artery dissection could reflect the severity of trauma, where multi-level fractures are associated with a higher risk of dissection.^20,21 It is possible that the early group in our study had a more severe cervical injury on admission, prompting urgent cricothyrotomy and subsequent conversion to tracheostomy.²² Additionally, patients with more extensive injury may be considered for early tracheostomy, given the higher likelihood of prolonged ventilator dependence. However, further research is warranted to investigate these hypotheses. Otherwise, for the level of cervical fracture associated with vertebral artery dissection, a prior review article found C5 to be the most common level associated with vertebral artery dissection occurring at 24.9 %, followed by C2 (19.6 %), C6 (19.6 %), C1 (11.6 %), C4 (10.9 %), C3 (7.0 %), and C7 (6.3 %).²³ In comparison, in our study, the early tracheostomy group was found to have more C3-C5 fractures, but no statistically significant differences were found among groups on analysis.

It is worth noting that for both open and percutaneous dilatational tracheostomy, an incision is commonly made in the inferior neck, which is in proximity to the C6-T1 ACDF incision site.²⁴ Creating a tracheostomy site adjacent to the surgical site could have raised concerns and delayed tracheostomy placement. Additionally, phrenic motoneurons are at the spinal cord level C3-C5. Fractures in this location could injure their associated descending tract of premotor neurons, resulting in virtually complete paralysis of the main inspiratory muscle, the diaphragm. In contrast, the C5-C8 injury level could contribute to the paralysis of remaining inspiratory and expiratory muscles, where the likelihood of breathing without mechanical intervention improves.^25,26 The greater respiratory impairment and complications in higher cervical injury may contribute to the need for early tracheostomy.^27,28 In our study, similar rates of acute respiratory failure were observed among the groups prior to the placement of tracheostomy.

Following adjusting for baseline comorbidities and surgical characteristics, the early group was found with a higher risk of 90-day pneumonia than the intermediate group. Sicker patients who had more acute presentation requiring early tracheostomy may have reduced respiratory function that predisposed them to increased risk of short-term pneumonia risk, although patients included in this study had similar CCI and SCI rates. In a prior database study investigating the timing of tracheostomy following traumatic concomitant cervical SCI, the early tracheostomy group (≤4 days) was associated with lower rates of respiratory complications, higher rates of ventilator-free days, intensive care unit-free days, and shorter hospital stays than late tracheostomy group (>4 days).²⁹ Similarly, a meta-analysis by Foran et al. found comparable respiratory benefits of early tracheostomy in acute spinal cord injury patients, with a particular focus on ventilator-associated pneumonia.⁸ In contrast, in the late tracheostomy group, no significant difference in the pneumonia rate was observed when compared to the early group. While current research highlights the potential short-term respiratory advantages of early tracheostomy, our findings suggest that further differences in pneumonia risk may emerge in the longer postoperative period, potentially reflecting variations in underlying patient characteristics.

Additionally, cerebrovascular accidents were found to be associated more with the early tracheostomy group than the late group. Previously, CVA has been reported in approximately 10 % of the patients who had a traumatic cerebrovascular injury, where high-energy trauma could disrupt the cervical carotid or vertebral artery wall, precipitating thromboembolism or occlusion of the artery.³⁰ Although vertebral artery dissection has been accounted for in the multivariate analysis, further investigation is needed to substantiate and validate this finding. For instance, the early tracheostomy group could have captured more CVA due to its proximity to the time of initial injury, which could introduce patients to a prothrombotic state.³¹ Ischemic stroke has been known to occur within 72 h of traumatic cerebrovascular injury,³² which may have been too early for them to be captured by the intermediate and late tracheostomy groups. Furthermore, the overall incidences of CVA were relatively low across all three tracheostomy groups (less than 10 events in each group), where the multivariate logistic regression model may be underpowered. Regarding incision and surgical-site complications, no significant variations were seen with deep wound infection or superficial wound infection. This observation is consistent with current research findings. While one may intuitively assume that the proximity of an early tracheostomy site and ACDF site could pose a risk for complication and infection, current literature suggests otherwise. Düsterwald et al. reported that patients undergoing anterior cervical surgery had similar outcomes between receiving tracheostomy the day of the surgery and 6–8 days later.¹⁰ Other studies also concur with these findings, where the cross-contamination risk between the ACDF site and the tracheostomy site is limited to 1 %.^11–13 These studies suggest early tracheostomy is safe in anterior cervical spine surgeries. On the other hand, Kelly et al. found that delayed tracheostomy (≥7 days) after operative cervical fixation, including both anterior and posterior approach, to be associated with increased ICU length of stay, ventilator days, and hospital length of stay in addition to increased morbidity.³³ Similar findings are also found in Anand et al.’s study comparing early and late tracheostomy following cervical SCI, as mentioned earlier.²⁹ Parallel to non-significant differences in surgical site infection, no significant difference was seen in the rate of dysphagia or anterior cervical spine reoperation rate, reaffirming the safety of early tracheostomy.

Given the nature of database research, this study has its inherent limitations. Given the non-granularity of the database study, we were unable to elicit the exact indications for early versus late tracheostomy and the types of tracheostomies, which could help elucidate some of our findings. Additionally, the tracheostomy time windows were selected based on previously reported average intervals to tracheostomy following surgical stabilization across multiple studies. While this approach provides a clinically grounded framework, it may introduce confounding variables, as timing decisions are influenced by heterogeneous clinical and institutional factors. Defining time windows based solely on objective criteria remains challenging in the context of variable real-world practices. Furthermore, we were not able to determine American Spinal Injuries Association (ASIA) scores, which could have provided valuable insights into the effect of the severity of injury on tracheostomy timing. Additional granular physiological data would also have been helpful in exploring pneumonia risk, which should be considered in future studies. Efforts have been made to address these limitations. Multiple covariates such as ventilator use, CCI, and SCI were used to reflect the severity of the injury. While these results may not be conclusive, these associative findings can be helpful in delineating patient characteristics and identifying pertinent surgical outcomes for hypothesis generation. The 90-day complications were calculated post-tracheostomy, as opposed to post-ACDF, which helped avoid confounding complications that precipitated tracheostomy, narrowing analysis to the post-tracheostomy window. Although this approach still has its own limitations from avoiding pre-tracheostomy input, arguably every study timeframe has its inherent strengths and limitations, particularly in retrospective analyses. Future studies should leverage patient-level datasets to enable advanced causal inference techniques such as propensity score matching and inverse probability weighting, which can better adjust for confounding factors. These approaches, combined with detailed clinical variables, could more accurately assess the impact of tracheostomy timing on outcomes.

5 Conclusion

The study offers valuable insights into the relationship between tracheostomy timing and postoperative outcomes following ACDF. In this retrospective cohort of patients with sub-axial cervical fractures undergoing ACDF, tracheostomy timing was not associated with increased surgical site infection or reoperation rates. Significant differences were observed across timing groups: prior to tracheostomy, the early group had a higher rate of vertebral artery dissection, while the late tracheostomy group had higher rates of COPD, CHF, and ventilator use. Intermediate tracheostomy was associated with a lower risk of postoperative pneumonia compared to early group. The late group was found with a lower rate of postoperative CVA than the early group. No significant differences in surgical site infection, dysphagia, or anterior reoperation rate were seen. These findings support the safety of early tracheostomy and highlight clinical differences across different tracheostomy timing cohorts.

CRediT authorship contribution statement

Hanzhi Yang: Methodology, Investigation, Formal analysis, Visualization, Writing – original draft. Jialun A. Chi: Methodology, Investigation, Data curation, Formal analysis, Visualization, Writing – review & editing. Li Jin: Visualization, Writing – review & editing. Jesse Wang: Writing – review & editing. Lawal Labaran: Writing – review & editing. Xudong Li: Conceptualization, Resources, Writing – review & editing, Supervision, Project administration.

Guardian/patient's consent

Due to the retrospective nature of the study, informed consent of the patients was not required because the study analyzed de-identified, anonymous clinical data.

Ethical statement

This study did not require ethical approval as it utilized de-identified patient data from the PearlDiver database, which does not contain personally identifiable information.

Data availability statement

The datasets generated during and/or analyzed during the current study are not publicly available due to PearlDiver database licensing agreement but are available from the corresponding author on reasonable request.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work the authors used ChatGPT 4o to improve the readability of the manuscript. After using this tool, the authors reviewed and edited the content as needed and took full responsibility for the content of the publication.

Funding statement

This research did not receive any specific grant from funding agencies in the public, commercial, or non-profit sectors.