Influencing factors for tracheostomy in patients with acute traumatic C3–C5 spinal cord injury and acute respiratory failure : Journal of the Chinese Medical Association

Secondary Logo

Journal Logo

Original Articles

Influencing factors for tracheostomy in patients with acute traumatic C3–C5 spinal cord injury and acute respiratory failure

Yu, Wen-Kuanga,b,c; Chen, Yu-Chuna,d; Chen, Wei-Chiha,c,d; Yi-Fong Su, Vincentc,e; Yang, Kuang-Yaoa,c,d,f; Kou, Yu Rub,d,*

Author Information
Journal of the Chinese Medical Association: February 2022 - Volume 85 - Issue 2 - p 167-174
doi: 10.1097/JCMA.0000000000000656
  • Open

Abstract

1. INTRODUCTION

Patients with acute traumatic cervical spinal cord injury (SCI), especially at a high level, have impaired major respiratory muscles and may develop acute respiratory failure (ARF) requiring invasive mechanical ventilation (IMV).1–3 A significant proportion of patients with high-level cervical SCI experience prolonged mechanical ventilation or become dependent on ventilators if the injury is severe.1–3 For these patients, tracheostomy has several advantages over translaryngeal intubation, such as enhanced pulmonary management and patient comfort.4,5 Additionally, compared with late tracheostomy, early tracheostomy (within 7 days following translaryngeal intubation) in patients with traumatic SCI may improve outcomes, including low rates of respiratory complications and in-hospital mortality, few ventilator days, and short intensive care unit (ICU) stay and hospital stay.6–10 However, tracheostomy may also lead to several intra- and postoperative complications, such as bleeding, infection, tracheal injury, and tracheal stenosis.4–6 Therefore, identifying patients with traumatic cervical SCI and ARF who need tracheostomy is of great clinical importance.

Several studies reported risk factors for tracheostomy in patients with traumatic cervical SCI, including being male, old age, high neurological level of injury, high injury severity score, low Glasgow Coma Scale (GCS), presence of other associated injury, and respiratory complications.10–16 However, these studies included patients with cervical SCI at all levels. The diaphragm is the major muscle responsible for breathing and is innervated by C3–C5. In general, patients with SCI at C2 and above develop complete paralysis of all respiratory muscles and are ventilator dependent;1,2 tracheostomy is necessary for improving long-term management in these patients. By contrast, patients with SCI at C6 and below have intact diaphragmatic function and usually can be weaned from mechanical ventilation,1,2 and thus the probability of undergoing tracheostomy is largely reduced in these patients. Patients with traumatic SCI at C3–C5 have varying degrees of diaphragmatic contraction impairment. Previous studies reported that the rates of tracheostomy in patients with traumatic SCI at C3–C5 and ARF have a wide range (27%-75%).10,14–20 The influencing factors for tracheostomy in this particular subset of patients remain unclear and making a decision about whether these patients undergo tracheostomy is challenging for clinicians.

In this retrospective case-control study, we aimed to analyze the influencing factors for predicting tracheostomy in patients with acute traumatic C3–C5 SCI and ARF requiring IMV, and these independent risk factors for tracheostomy were identified through univariate and multivariate logistic regression analyses. Receiver operating characteristic (ROC) curve analyses were performed to assess the predictive performance of independent influencing factors for tracheostomy.

2. METHODS

2.1. Study population

This retrospective observational study was conducted at Taipei Veterans General Hospital, Taiwan, a tertiary medical center with 3000 beds, and was approved by the Institutional Ethical Review Board (IRB) of Taipei Veterans General Hospital (VGHTPE-IRB No. 201702003C). The requirement for informed consent was waived by the IRB according to institutional guidelines for retrospective observational studies. Between January 2007 and December 2016, patients newly diagnosed with acute traumatic C3–C5 SCI and ARF requiring initial endotracheal intubation and IMV were identified. The exclusion criteria were as follows: (1) duration between the onset of traumatic C3–C5 SCI and ARF was longer than 7 days, (2) duration of IMV was shorter than 48 hours, (3) passed away within 7 days after hospitalization, (4) transferred to other hospitals for family requests or for the isolation of active pulmonary mycobacteria tuberculosis infection; (5) age of <18 years; and (6) incomplete data. In our traumatic cervical SCI patients, the decision for the need for tracheostomy was considered after an initial period (generally the first 7 days) of stabilization on IMV.6–10 However, the treatment of tracheostomy was made only after obtaining the informed consent from patients or their surrogates. The enrolled patients were divided into two study groups. The no-tracheostomy group included patients who were extubated directly without tracheostomy, whereas the tracheostomy group included patients who underwent tracheostomy during weaning from IMV.

2.2. Definitions

The neurological level of cervical SCI was determined according to the most caudal segment of the cervical SCI with normal motor and sensory function on both sides of the body.21 Acute respiratory distress syndrome (ARDS) was assessed according to the Berlin definition based on blood gases, chest radiographs, and clinical data.22 Acute kidney injury (AKI) was defined using serum creatinine levels and urine outputs in accordance with the KDIGO clinical practice guideline for AKI.23 Briefly, AKI was diagnosed as an increase of ≥0.3 mg/dL in serum creatinine level within 48 hours or ≥50% times the baseline. The presence of shock was defined as a systolic blood pressure of <90 mmHg despite adequate fluid resuscitation along with vasopressors for more than 48 hours.24,25 Weaning parameters, including rapid shallow breathing index (RSBI), maximal inspiratory pressure (Pimax) and maximal expiratory pressure (Pemax), were measured when cervical SCI patients started to wean from the IMV. RSBI was calculated as the ratio between respiratory frequency and tidal volume (breaths/min/L), which were measured and recorded using a handheld respirometer during spontaneous breathing trial without any ventilator assistance.26 Pimax and Pemax were defined as the maximum pressure generated during inspiration and expiration, respectively, against an occluded airway.26

2.3. Weaning protocol

The survey of weaning parameters was initiated only when patients achieved a satisfactory progress as judged by physicians. The weaning protocol began in the morning and the criteria for weaning from the ventilation, start/interruption of spontaneous breathing trial, and extubation/reintubation were reported previously.27 Various weaning parameters, including Pimax, Pemax, minute ventilation (Ve) and RSBI were measured using Wright Respirometer Haloscale Standard (nSpire Health Ltd., Hertford, United Kingdom) and Hand-held Inspiratory Force Meter (Boehringer, Ingelheim, DE) with brief discontinuation of mechanical ventilation support. After obtaining these data, the patient’s mechanical ventilation was quickly resumed. The weaning parameters were measured at least one time and, in the cases of more than one time, the data obtained from the first time were used. When patients were judged to be ready for weaning, they were subjected to spontaneous breathing trial using T-piece with the same FiO2 from a large-volume jet nebulizer for 1 hour. When the patients successfully completed the spontaneous breathing trial, they were extubated and given a nasal cannula or air-entrainment mask for oxygen supplement to maintain satisfactory oxygenation. In patients who were not able to complete the spontaneous breathing trial, subsequent attempts were made a few days later when their clinical conditions were allowed. Successful extubation was defined as patients free from the IMV for more than 48 hours after extubation. Reinstitution of noninvasive mechanical ventilation or reintubation with IMV support within 48 hours was considered an extubation failure. Patients in this study did not receive any rehabilitation program for weaning and received usual neurocritical care, including the bundle to prevent ventilator-associated pneumonia (VAP) and sedation or pain control.

2.4. Data collection

The data used in this retrospective study were collected from medical charts and electronic medical records and included gender, age, body mass index, smoking history, Charlson comorbidity index,28 laboratory values, conscious levels as defined by the GCS,29 and vital signs at admission. The weaning parameters measured during weaning from IMV were Pimax, Pemax, Ve, and RSBI. Additionally, treatment complications and clinical outcomes, including VAP, ARDS, shock, bacteremia, urinary tract infection (UTI), AKI, length of ICU stay, length of hospitalized stay, and mortality, were collected.

2.5. Statistical analysis

We analyzed differences in data between the no-tracheostomy and tracheostomy groups. The Kolmogorov–Smirnov test was used in checking the distribution of the continuous variables, which were presented as medians with interquartile range (IQR) and analyzed with the Mann-Whitney U test. Categorical variables were presented as numbers and percentages and analyzed with Chi-square test or Fisher’s exact test. Variables associated with tracheostomy (p < 0.2) in univariate analysis were subjected to Spearman rank correlation test to determine their correlations. If two univariate factors in the multivariate analysis were highly associated, then only one factor was selected for analysis to avoid collinearity. Afterwards, multivariate logistic regression analysis was performed to identify the independent risk factors associated with tracheostomy. Odds ratio (OR) and OR after adjustment for confounding factors with 95% confidence interval (CI) were calculated. The performance of variables for predicting tracheostomy or no tracheostomy was determined through ROC curve analysis, and areas under the ROC curves (AUC) were calculated. A two-tail p-value of less than 0.05 was considered significant. All analyses were performed using SPSS version 19.0 (IBM, Armonk, NY).

3. RESULTS

3.1. Characteristics of patients

During the 10-year study period, 132 patients were admitted to Taipei Veterans General Hospital had a diagnosis of acute traumatic C3–C5 SCI and subsequently developed ARF requiring initial endotracheal intubation and IMV for more than 48 hours. A total of 21 patients were excluded according to the exclusion criteria (Fig. 1). The remaining 101 patients had a median age of 60 (IQR, 44-74) were enrolled. Among the 101 patients enrolled, 89 (88.1%) patients received computed tomography scan and 92 (91.1%) patients received magnetic resonance imaging for the diagnosis of traumatic C3-C5 SCI. The numbers of enrolled patients with C3, C4, and C5 injury levels were 35 (34.6%), 43 (42.6%), and 23 (22.8%), respectively. Among the 101 patients enrolled, 59 patients (58.4%) were extubated directly without tracheostomy (no-tracheostomy group), whereas 42 patients (41.6%) underwent tracheostomy during weaning from IMV (tracheostomy group; Fig. 1). In the tracheostomy group, 15 patients (35.7%) experienced extubation failure. The timing of performing tracheotomy following hospital admission and initial translaryngeal intubation was 28 days (IQR, 19-40 days) and 25 days (IQR, 17-40 days), respectively. As a whole group, the duration between initiation of IMV due to ARF and the first attempt of weaning from IMV (the time of measuring initial weaning parameters) was 6 days (IQR, 3-14 days) When the patients were discharged from the hospital, two patients in the no-tracheostomy group required noninvasive positive pressure ventilation, and 19 patients in the tracheostomy group were ventilator dependent. A total of 80 patients (79.2%) were successfully liberated from IMV (Fig. 1).

F1
Fig. 1:
Flowchart of the study population. ARF = acute respiratory failure; IMV = invasive mechanical ventilation; NIPPV = noninvasive positive pressure ventilators; SCI = spinal cord injury; TB = tuberculosis.

3.2. Demographic and baseline clinical data

Table 1 shows the comparison between the groups in terms of demographic characteristics and baseline clinical variables measured at admission. Compared with the patients in the no-tracheostomy group, the patients in the tracheostomy group had a higher proportion of C3 level injury (p = 0.010), lower level of whole blood hemoglobin (p < 0.001), and lower level of GCS (p = 0.001). No significant differences in demographic characteristics and other baseline clinical variables were observed between the groups (Table 1). Fig. 2 shows the comparisons of four initial weaning parameters between the groups. Compared with the patients in the no-tracheostomy group, the patients in the tracheostomy group had significantly lower levels of Pimax (p < 0.001, Fig. 2A), Pemax (p < 0.001, Fig. 2B), and Ve (p = 0.014, Fig. 2C) but had a significantly higher level of RSBI (p < 0.001, Fig. 2D).

Table 1 - Demographic characteristics and baseline clinical data of the two study groups
No tracheostomy (N = 59) Tracheostomy (N = 42) p
Age (years) 58 (44-72) 60 (42-76) 0.644
Male sex 52 (88) 36 (86) 0.769
BMI (kg/m2) 24.1 (21.7-26.1) 23.8 (20.9-25.4) 0.341
Smoking 18 (31) 16 (38) 0.522
Causes of injury 0.347
 Fall 28 (47.5) 17 (40.5)
 Traffic accidents 30 (50.8) 22 (52.4)
 Others 1 (1.7) 3 (7.1)
C3 level injury 14 (24) 21 (50) 0.010
Co-morbidities
 Diabetes mellitus 14 (24) 9 (21) 0.815
 Hypertension 19 (32) 16 (38) 0.672
 Coronary artery disease 5 (8.5) 5 (11.9) 0.738
 Chronic airway diseases 4 (6.8) 4 (9.5) 0.716
 Liver cirrhosis 0 (0) 1 (2.4) 0.416
 Stroke 2 (3.4) 1 (2.4) 1.000
 Uremia 1 (1.7) 0 (0) 1.000
 Malignancy 2 (3.4) 1 (2.4) 1.000
Charlson comorbidity index 1 (0-2) 1 (0-2) 0.679
At admission
 GCS 15 (9-15) 10 (6-11) 0.001
 TBI 6 (10.2) 7 (16.7) 0.377
 Chest trauma 4 (6.8) 6 (14.3) 0.312
 Mean BP (mmHg) 89 (78-97) 83 (69-94) 0.097
 BUN (mg/dL) 19 (17-24) 20 (17-26) 0.712
 Creatinine (mg/dL) 0.97 (0.78-1.21) 0.90 (0.66-1.30) 0.486
 Sodium (mmol/L) 139 (137-140) 138 (134-142) 0.335
 Potassium (mmol/L) 4.0 (3.7-4.2) 4.0 (3.8-4.3) 0.374
 WBC (103/mm3) 11.1 (8.2-13.0) 10.5 (8.5-12.8) 0.785
 HGB (g/dL) 12.3 (11.1-13.4) 10.8 (9.7-12.2) <0.001
 Platelet (103/mm3) 189 (147-222) 186 (144-229) 0.981
Treatment
 Surgery 50 (84.7) 34 (81.0) 0.788
 Steroid use 37 (62.7) 25 (59.5) 0.836
Chronic airway diseases include asthma and chronic obstructive pulmonary disease.
Chest trauma includes pulmonary contusion, rib fractures, hemothorax and pneumothorax.
Mean BP is defined as 1/3 systolic BP plus 2/3 diastolic BP.
Continuous data are expressed as median with interquartile range (IQR) and were compared by Mann-Whitney U test. Categorical variables are expressed as number of patients (%) and were compared by Chi-square test or Fisher’s exact test. A p-value of <0.05 was considered statistically significant.
BMI = body mass index; BP = blood pressure; BUN = blood urea nitrogen; GCS = Glasgow coma scale; HGB = hemoglobin; TBI = traumatic brain injury; WBC = white blood cell.

F2
Fig. 2:
Comparisons of the four weaning parameters measured at the first weaning attempt between the two study groups. Parameters: maximal inspiratory pressure (Pimax; A), maximal expiratory pressure (Pemax; B), minute ventilation (Ve; C), and rapid shallow breathing index (RSBI; D). The no-tracheostomy group included patients who were extubated directly without tracheostomy. The tracheostomy group included patients who underwent tracheostomy during weaning from invasive mechanical ventilation. Data were compared with Mann-Whitney U test and are presented as boxplots. A p-value of <0.05 was considered statistically significant.

3.3. Clinical complications and hospital outcomes

Table 2 shows the comparisons of clinical complications and hospital outcomes between the groups. Compared with the patients in the no-tracheostomy group, the patients in the tracheostomy group had higher rates of VAP (p < 0.001), bacteremia (p = 0.027), UTI (p = 0.010), and AKI (p = 0.013) and longer durations of IMV (p < 0.001), ICU stay (p < 0.001), and hospitalization (p < 0.001). However, no significant differences in the incidences of ARDS and shock were observed between the groups. Four patients (10%) in the tracheostomy group and one patient (2%) in the no-tracheostomy group died during hospitalization. The timing of tracheostomy following translayrngeal intubation was not different from whether complications develop or not, including VAP (p = 0.818), ARDS (p = 0.124), shock (p = 0.771), bacteremia (p = 0.284), UTI (p = 0.635), and AKI (p = 0.258).

Table 2 - Comparisons of clinical complications and hospital outcomes between the two study groups
No tracheostomy (N = 59) Tracheostomy (N = 42) p
Complications
 VAP 21 (35.6) 31 (73.8) <0.001
 ARDS 3 (5.1) 5 (11.9) 0.272
 Shock 22 (37.3) 23 (54.8) 0.105
 Bacteremia 8 (13.6) 15 (33.3) 0.027
 UTI 14 (23.7) 21 (50.0) 0.010
 AKI 16 (27.1) 22 (52.4) 0.013
Outcomes
 IMV days 9 (5-22) 52 (34-69) <0.001
 ICU days 17 (9-30) 53 (33-72) <0.001
 Hospitalization days 44 (38-55) 73 (55-95) <0.001
 Died in hospital 1 (1.7) 4 (9.5) 0.157
Continuous data are expressed as median with interquartile range (IQR) and were compared by Mann-Whitney U test. Categorical variables are expressed as number of patients (%) and were compared by Chi-square test or Fisher’s exact test. A p-value of <0.05 was considered statistically significant.
AKI = acute kidney injury; ARDS = acute respiratory distress syndrome; ICU = intensive care unit; IMV = invasive mechanical ventilation; UTI = urinary tract infection; VAP = ventilator-associated pneumonia.

3.4. Independent prognostic factors for tracheostomy

Table 3 shows the results of univariate and multivariate logistic regression analyses for identifying the prognostic factors for tracheostomy. The univariate analyses revealed that C3 level injury (p = 0.007), whole blood hemoglobin levels (p = 0.001), GCS at admission (p = 0.001), Pimax (p < 0.001), Pemax (p = 0.003), Ve (p = 0.022), RSBI (p < 0.001), VAP (p < 0.001), bacteremia (p = 0.021), and AKI (p = 0.011) were significantly associated with tracheostomy. Results from the Spearman rank correlation analyses of collinearity between any two continuous variables, including whole blood hemoglobin levels, GCS at admission, Pimax, Pemax, Ve and RSBI are shown in Table S1, https://links.lww.com/JCMA/A123. As shown, HGB at admission was highly correlated with GCS admission (p < 0.001). Initial Pimax was highly correlated with initial Pemax (p < 0.001) and initial RSBI (p < 0.001). Therefore, in the multivariate logistic regression analysis, these two variables (HGB at admission and initial Pimax) were not selected as cofactors to avoid collinearity. After adjustment for other cofactors, the initial RSBI (p < 0.001) remained an independent risk factor, with significantly increased adjusted ORs for tracheostomy, whereas GCS at admission (p = 0.003) was an independent preventive factor, with a significant decrease in adjusted ORs for tracheostomy (Table 3).

Table 3 - Univariate and multivariate logistic regression analyses of influencing factors for tracheostomy
Univariate Multivariate
OR (95% CI) p aOR (95% CI) p
C3 level injury 3.21 (1.37-7.53) 0.007 1.18 (0.32-4.29) 0.803
HGB at admission 0.63 (0.48-0.82) 0.001 a
GCS at admission 0.84 (0.75-0.093) 0.001 0.76 (0.64-0.91) 0.003
VAP 5.10 (2.14-12.17) <0.001 4.77 (0.97-17.11) 0.056
Bacteremia 3.19 (1.19-8.52) 0.021 1.02 (0.20-5.15) 0.986
AKI 2.96 (1.28-6.81) 0.011 2.21 (0.62-7.94) 0.224
Initial Pimax 0.92 (0.88-0.96) <0.001 a
Initial Pemax 0.96 (0.94-0.99) 0.003 0.98 (0.95-1.01) 0.128
Initial Ve 0.83 (0.70-0.97) 0.022 0.88 (0.69-1.14) 0.344
Initial RSBI 1.04 (1.02-1.05) <0.001 1.04 (1.02-1.06) <0.001
A p-value of <0.05 was considered statistically significant.
AKI = acute kidney injury; aOR = odds ratio after adjustment for other confounding factors; CI = confidence interval; GCS = Glasgow coma scale; HGB = haemoglobin; OR = odds ratio; Pemax = maximal expiratory pressure; Pimax = maximal inspiratory pressure; RSBI = rapid shallow breathing index; VAP = ventilator-associated pneumonia; Ve = minute ventilation.
aHGB at admission and initial Pimax were not selected as cofactors in the multivariate logistic regression analysis to avoid collinearity.

3.5. ROC curve analyses for predicting tracheostomy or no tracheostomy

Given that GCS at admission and initial RSBI were found to be independent influencing factors for tracheostomy, the predictive performance of the two factors was assessed through ROC curve analyses (Fig. 3). GCS at admission (AUC = 0.69, 95% CI 0.58-0.79, p = 0.001) had a moderate discriminative performance for predicting no tracheostomy (Fig. 3A), whereas initial RSBI (AUC = 0.82, 95% CI 0.74-0.91, p < 0.001) had excellent discriminative performance for predicting tracheostomy (Fig. 3B).

F3
Fig. 3:
Receiver-operating characteristic curve analyses of Glasgow Coma Scale (GCS; A) at admission and initial rapid shallow breathing index (RSBI; B). GCS was for prediction of no tracheostomy, whereas RSBI was for the prediction of tracheostomy. AUC = area under the curve.

3.6. Synergistic impact of the two influencing factors on tracheostomy

According to the ROC curve analyses, the cutoff values of GCS at admission and initial RSBI were 12.5 and 98.1, respectively (Fig. 3A, B). We further analyzed the possible synergistic impacts of the prognostic factors on tracheostomy. The study patients were divided into four subgroups (Fig. 4). Patients with a GCS of ≥13 at admission and initial RSBI of <98.1 were included in the reference group (subgroup 1). When the patients had a GCS of <13 at admission and initial RSBI of <98.1 (subgroup 2; OR = 8.57, 95% CI 2.22-33.11, p < 0.001) or had a GCS of ≥13 at admission and initial RSBI of ≥98.1 (subgroup 3; OR = 60.00, 95% CI 5.18 - 694.51, p < 0.001), the OR for tracheostomy significantly increased. Remarkably, when patients had a GCS of <13 at admission and initial RSBI of ≥98.1, the OR for tracheostomy synergistically increased (subgroup 4; OR = 228.00; 95% CI 22.17-2344.34, p < 0.001).

F4
Fig. 4:
Analysis of the odds ratio (OR) for tracheostomy across four subgroups of patients who were stratified by Glasgow Coma Scale (GCS) score at admission and initial rapid shallow breathing index (RSBI). The cutoff values of GCS at admission and initial RSBI to define the subgroups were obtained from receiver-operating characteristic curve analyses. CI = confidence interval.

4. DISCUSSION

Tracheostomy confers various clinical benefits and improves the clinical outcomes of patients with traumatic cervical SCI4–10 but bears certain risks of intra- and postoperative complications.4–6 Therefore, investigating prognostic factors related to tracheostomy among patients with traumatic cervical SCI is of clinical importance. In this study, we conducted an investigation to address this issue in a particular subset of patients with acute traumatic SCI at levels of C3–C5. We found that the patients in the tracheostomy group exhibited several clinical features that were significantly distinct from those in the patients in the no-tracheostomy group, as revealed by univariate analyses. However, multivariate logistic regression analysis revealed that low GCS at admission and high RSBI at the first weaning attempt were the only two independent risk factors for tracheostomy. On the basis of the cutoff values obtained from the ROC curve analyses, we further analyzed the possible synergistic impacts of these influencing factors on tracheostomy by stratifying the patients into four subgroups. Remarkably, these influencing factors seemed to act synergistically, and OR for tracheostomy increased greatly in patients with both influencing factors, as compared to the OR in patients with either influencing factor.

Several studies analyzed data from patients with traumatic cervical SCI at all levels and reported risk factors for tracheostomy.10–16 A recent meta-analysis11 including 16 studies containing 9697 patients indicated that being male, high neurological level of injury, high injury severity score, low GCS, thoracic injury, and respiratory complications are risk factors for tracheostomy. Other studies showed additional risk factors, including high age,13,15 complete cervical SCI,10,13,16 intubation on arrival,12,13,16 and presence of other associated injuries.12,16 However, necessities of tracheostomy are relatively well defined in patients with certain neurological levels of injury. Patients with traumatic SCI at C2 and above are usually ventilator-dependent because of the paralysis of all respiratory muscles and require tracheostomy for long-term management.1,2 Patients with traumatic SCI at C6 and below can usually be weaned from the ventilator because of intact diaphragmatic function and do not need tracheostomy unless they have other major complications.1,2 Making a decision about whether patients with traumatic SCI at C3–C5 need tracheostomy is challenging because these patients have varying degrees of impairment of diaphragmatic contraction and the risk factors for this particular subset of patients remain unclear. Indeed, the rates of tracheostomy in patients with traumatic SCI at C3–C5 and ARF have a wide range (27%-75%).10,14–20 Our findings regarding low GCS at admission and high RSBI as two independent factors may help clinicians decide whether to perform tracheostomy on this particular subset of patients.

In this study, patients in the tracheostomy group had a significantly lower GCS at admission. One possible reason for low GCS in traumatic SCI patient is a co-occurring traumatic brain injury.30 However, the incidence of traumatic brain injury in the groups did not differ significantly. The other possible reason is the impairment of motor responses to painful stimuli because of complete or incomplete SCI. A recent meta-analysis11 including four studies containing 5215 patients with cervical SCI at all levels indicated that patients with GCS of <8 had a pooled OR of 6.03 for tracheostomy. Patients with GCS of ≥8 was used as the reference group. Thus, low GCS is also a risk factor for tracheostomy in patients with traumatic SCI at C3–C5 and ARF. RSBI is a widely used weaning predictor during spontaneous breathing trial in ICU.29 Although the cutoff values of RSBI for predicting weaning success vary among patients with different etiologies, patients with an RSBI greater than the cutoff value are known to have a high risk of weaning failure.31 Hence, patients with cervical SCI usually display a rapid shallow breathing pattern because of the paralysis of major respiratory muscles.32 In this study, the patients in the tracheostomy group had a significantly higher RSBI at the first weaning attempt compared with the patients in the no tracheostomy group. The higher RSBI may imply that these patients had an unbalance between the increased work of breath and the decreased capacity of respiratory muscle pump at that time and are vulnerable to weaning failure. Both conditions may contribute to the possibility of prolonged IMV and need of tracheostomy. Our findings showed that initial RSBI is a novel influencing factor, providing an additional consideration when determining who merits tracheostomy in this subset of patients. Additionally, our findings regarding the synergistic effect of these two influencing factors for tracheostomy may help ICU physicians in determining patients who considerably need tracheostomy, that is, patients who have both risk factors.

Previous studies reported that early tracheostomy may improve several clinical outcomes in patients with traumatic SCI.6–10 The majority of these studies performed an early tracheostomy within 7 days from either time of injury or tracheal intubation.6,7,9,10 Taken together, these results suggested that our patients underwent tracheostomy at the stage of failure-to-wean and prolonged IMV, which is commonly defined as more than 21 days of IMV.24 Indeed, we found that the tracheostomy group had more complications and worse clinical outcomes, that is, patients with tracheotomy had relatively unfavorable conditions. In traumatic cervical SCI patients, tracheostomy can be considered after an initial period (generally the first 7 days) of stabilization on IMV.6–10 However, a practical decision can be made only after the informed consent from patients themselves or their surrogates was obtained. In this context, tracheostomy is difficult for patients or their surrogates to accept in Taiwan, particularly for older people, because of the misconceptions and fear about this procedure.33 The reason for the delayed timing of tracheostomy in our study was the requests of the patients’ families to delay or avoid tracheostomy. Perhaps, the risk factors we identified may be helpful in tracheostomy decision-making communication34 among patients with traumatic SCI at C3–C5 and in promoting early tracheostomy that hopefully may improve clinical outcomes.

The strength of this study is that, to our knowledge, this is the first investigation that focused on traumatic C3-C5 SCI patients with respiratory failure to investigate the influencing factors associated with tracheostomy. However, the current study had some limitations. First, this study was a retrospective investigation at a single-center and involved a relatively small sample size. Thus, future multicenter investigations with larger sample sizes are warranted to confirm our findings. Second, many data about the serial change of conscious level as assessed by Glasgow Soma Scale, the severity of initial injury as defined by Injury Severity Score, the American Spinal Injury Association Impairment Scale,21 the Acute Physiology and Chronic Health Evaluation score,35 the Simplified Acute Physiology Score,36 and the presence or absence of a nerve root injury, were not fully recorded from the medical chart. Third, our hospital is a tertiary referral center primarily for the care of military veterans. Therefore, the proportion of males with old age was higher. Finally, given that the data were collected over a 10-year period, consistently similar treatment strategies might not have been used to treat the study population.

In conclusion, this study demonstrated that low GCS at admission and high initial RSBI are the two independent influencing factors that synergistically impact on tracheostomy in patients with acute traumatic C3–C5 SCI and ARF requiring IMV. These findings may help clinicians to decide whether to perform tracheostomy on this subset of patient population.

ACKNOWLEDGMENTS

The authors are grateful to the KGSupport Academic Submission Services for assisting with language editing. This research was funded by grants from Taipei Veterans General Hospital—National Yang-Ming University Excellent Scientist Cultivation Program (No.107-V-A-001, W.-K.Y.) and the Ministry of Science and Technology, Taiwan (MOST 107-2320-B-010-027-MY3, Y.R.K.).

APPENDIX A. SUPPLEMENTARY DATA

Supplementary data related to this article can be found at https://links.lww.com/JCMA/A123.

REFERENCES

1. Sheel AW, Welch JF, Townson AF. Respiratory management following spinal cord injury. Eng JJ, Teasell RW, Miller WC, Wolfe DL, Townson AF, Hsieh JTC, et al, editors. In: Spinal cord injury rehabilitation evidence version 6.0. Vancouver; 2018, p. 172.
2. Arora S, Flower O, Murray NP, Lee BB. Respiratory care of patients with cervical spinal cord injury: a review. Crit Care Resusc. 2012;14:64–73.
3. Como JJ, Sutton ER, McCunn M, Dutton RP, Johnson SB, Aarabi B, et al. Characterizing the need for mechanical ventilation following cervical spinal cord injury with neurologic deficit. J Trauma. 2005;59:912–6.
4. Ganuza JR, Oliviero A. Tracheostomy in spinal cord injured patients. Transl Med Unisa. 2011;1:151–72.
5. Galeiras Vázquez R, Rascado Sedes P, Mourelo Fariña M, Montoto Marqués A, Ferreiro Velasco ME. Respiratory management in the patient with spinal cord injury. Biomed Res Int. 2013;2013:168757.
6. Romero J, Vari A, Gambarrutta C, Oliviero A. Tracheostomy timing in traumatic spinal cord injury. Eur Spine J. 2009;18:1452–7.
7. Mubashir T, Arif AA, Ernest P, Maroufy V, Chaudhry R, Balogh J, et al. Early versus late tracheostomy in patients with acute traumatic spinal cord injury: a systematic review and meta-analysis. Anesth Analg. 2021;132:384–94.
8. Anand T, Hanna K, Kulvatunyou N, Zeeshan M, Ditillo M, Castanon L, et al. Time to tracheostomy impacts overall outcomes in patients with cervical spinal cord injury. J Trauma Acute Care Surg. 2020;89:358–64.
9. Flanagan CD, Childs BR, Moore TA, Vallier HA. Early tracheostomy in patients with traumatic cervical spinal cord injury appears safe and may improve outcomes. Spine (Phila Pa 1976). 2018;43:1110–6.
10. Leelapattana P, Fleming JC, Gurr KR, Bailey SI, Parry N, Bailey CS. Predicting the need for tracheostomy in patients with cervical spinal cord injury. J Trauma Acute Care Surg. 2012;73:880–4.
11. Wang Y, Guo Z, Fan D, Lu H, Xie D, Zhang D, et al. A meta-analysis of the influencing factors for tracheostomy after cervical spinal cord injury. Biomed Res Int. 2018;2018:5895830.
12. Higashi T, Eguchi H, Wakayama Y, Sumi M, Saito T, Inaba Y. Analysis of the risk factors for tracheostomy and decannulation after traumatic cervical spinal cord injury in an aging population. Spinal Cord. 2019;57:843–9.
13. Lee DS, Park CM, Carriere KC, Ahn J. Classification and regression tree model for predicting tracheostomy in patients with traumatic cervical spinal cord injury. Eur Spine J. 2017;26:2333–9.
14. Childs BR, Moore TA, Como JJ, Vallier HA. American spinal injury association impairment scale predicts the need for tracheostomy after cervical spinal cord injury. Spine (Phila Pa 1976). 2015;40:1407–13.
15. Yugué I, Okada S, Ueta T, Maeda T, Mori E, Kawano O, et al. Analysis of the risk factors for tracheostomy in traumatic cervical spinal cord injury. Spine (Phila Pa 1976). 2012;37:E1633–8.
16. Branco BC, Plurad D, Green DJ, Inaba K, Lam L, Cestero R, et al. Incidence and clinical predictors for tracheostomy after cervical spinal cord injury: a National Trauma Databank review. J Trauma. 2011;70:111–5.
17. Jeong TS, Lee SG, Kim WK, Ahn Y, Son S. Predictive values of magnetic resonance imaging features for tracheostomy in traumatic cervical spinal cord injury. J Korean Neurosurg Soc. 2018;61:582–91.
18. Beom JY, Seo HY. The need for early tracheostomy in patients with traumatic cervical cord injury. Clin Orthop Surg. 2018;10:191–6.
19. Ganuza JR, Garcia Forcada A, Gambarrutta C, Diez De La Lastra Buigues E, Merlo Gonzalez VE, Paz Fuentes F, et al. Effect of technique and timing of tracheostomy in patients with acute traumatic spinal cord injury undergoing mechanical ventilation. J Spinal Cord Med. 2011;34:76–84.
20. Berney SC, Gordon IR, Opdam HI, Denehy L. A classification and regression tree to assist clinical decision making in airway management for patients with cervical spinal cord injury. Spinal Cord. 2011;49:244–50.
21. Kirshblum SC, Waring W, Biering-Sorensen F, Burns SP, Johansen M, Schmidt-Read M, et al. Reference for the 2011 revision of the International Standards for Neurological Classification of Spinal Cord Injury. J Spinal Cord Med. 2011;34:547–54.
22. Ranieri VM, Rubenfeld GD, Thompson BT, Ferguson ND, Caldwell E, Fan E, et al.; ARDS Definition Task Force. Acute respiratory distress syndrome: the Berlin Definition. JAMA. 2012;307:2526–33.
23. KDIGO AKI Work Group: KDIGO clinical practice guideline for acute kidney injury. Kidney Int Suppl. 2012;2:1138.
24. Pan SW, Kao HK, Lien TC, Chen YW, Kou YR, Wang JH. Acute kidney injury on ventilator initiation day independently predicts prolonged mechanical ventilation in intensive care unit patients. J Crit Care. 2011;26:586–92.
25. Yu WK, Ko HK, Ho LI, Wang JH, Kou YR. Synergistic impact of acute kidney injury and high level of cervical spinal cord injury on the weaning outcome of patients with acute traumatic cervical spinal cord injury. Injury. 2015;46:1317–23.
26. Yang KL, Tobin MJ. A prospective study of indexes predicting the outcome of trials of weaning from mechanical ventilation. N Engl J Med. 1991;324:1445–50.
27. Bien MY, Shui Lin Y, Shih CH, Yang YL, Lin HW, Bai KJ, et al. Comparisons of predictive performance of breathing pattern variability measured during T-piece, automatic tube compensation, and pressure support ventilation for weaning intensive care unit patients from mechanical ventilation. Crit Care Med. 2011;39:2253–62.
28. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373–83.
29. Teasdale G, Jennett B. Assessment of coma and impaired consciousness. A practical scale. Lancet. 1974;2:81–4.
30. Macciocchi S, Seel RT, Thompson N, Byams R, Bowman B. Spinal cord injury and co-occurring traumatic brain injury: assessment and incidence. Arch Phys Med Rehabil. 2008;89:1350–7.
31. Karthika M, Al Enezi FA, Pillai LV, Arabi YM. Rapid shallow breathing index. Ann Thorac Med. 2016;11:167–76.
32. Denton M, McKinlay J. Cervical cord injury and critical care. Continuing Educ Anaesth Crit Care Pain. 2009;9:826.
33. Chen HC, Song L, Chang HC, Hsu MT. Factors related to tracheostomy timing and ventilator weaning: findings from a population in Northern Taiwan. Clin Respir J. 2018;12:97–104.
34. Huang CH, Chen IH. The clinical application of the tracheostomy decision-making program in respiratory care center prolonged mechanical ventilation patients in Taiwan. Int J Gen Med. 2020;13:1487–94.
35. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: a severity of disease classification system. Crit Care Med. 1985;13:818–29.
36. Le Gall JR, Lemeshow S, Saulnier F. A new Simplified Acute Physiology Score (SAPS II) based on a European/North American multicenter study. JAMA. 1993;270:2957–63.
Keywords:

Acute respiratory failure; Acute traumatic cervical spinal cord injury; Glasgow Coma Scale; Rapid shallow breathing index; Tracheostomy

Supplemental Digital Content

Copyright © 2022, the Chinese Medical Association.