This randomized clinical trial investigates if the surgical treatment of unstable chest wall injuries improves ventilator-free days and outcomes compared with nonoperative treatment.
Key Points
Question
Does surgical treatment of unstable chest wall injuries improve ventilator-free days and outcomes compared to nonoperative treatment?
Findings
In this randomized clinical trial, 207 patients were randomly assigned to receive operative or nonoperative treatment, and surgery was associated with improvement in ventilator-free days and decreased length of hospitalization in the subgroup of patients who were receiving mechanical ventilation at the time of randomization. Additionally, there was decreased mortality in the surgery group (0% vs 6%; P = .01).
Meanings
Patients who are receiving mechanical ventilation may benefit from surgical treatment of unstable chest wall injuries, but nonventilated patients demonstrated little benefit from surgical intervention.
Abstract
Importance
Unstable chest wall injuries have high rates of mortality and morbidity. In the last decade, multiple studies have reported improved outcomes with operative compared with nonoperative treatment. However, to date, an adequately powered, randomized clinical trial to support operative treatment has been lacking.
Objective
To compare outcomes of surgical treatment of acute unstable chest wall injuries with nonsurgical management.
Design, Setting, and Participants
This was a multicenter, prospective, randomized clinical trial conducted from October 10, 2011, to October 2, 2019, across 15 sites in Canada and the US. Inclusion criteria were patients between the ages of 16 to 85 years with displaced rib fractures with a flail chest or non–flail chest injuries with severe chest wall deformity. Exclusion criteria included patients with significant other injuries that would otherwise require prolonged mechanical ventilation, those medically unfit for surgery, or those who were randomly assigned to study groups after 72 hours of injury. Data were analyzed from March 20, 2019, to March 5, 2021.
Interventions
Patients were randomized 1:1 to receive operative treatment with plate and screws or nonoperative treatment.
Main Outcomes and Measures
The primary outcome was ventilator-free days (VFDs) in the first 28 days after injury. Secondary outcomes included mortality, length of hospital stay, intensive care unit stay, and rates of complications (pneumonia, ventilator-associated pneumonia, sepsis, tracheostomy).
Results
A total of 207 patients were included in the analysis (operative group: 108 patients [52.2%]; mean [SD] age, 52.9 [13.5] years; 81 male [75%]; nonoperative group: 99 patients [47.8%]; mean [SD] age, 53.2 [14.3] years; 75 male [76%]). Mean (SD) VFDs were 22.7 (7.5) days for the operative group and 20.6 (9.7) days for the nonoperative group (mean difference, 2.1 days; 95% CI, −0.3 to 4.5 days; P = .09). Mortality was significantly higher in the nonoperative group (6 [6%]) than in the operative group (0%; P = .01). Rates of complications and length of stay were similar between groups. Subgroup analysis of patients who were mechanically ventilated at the time of randomization demonstrated a mean difference of 2.8 (95% CI, 0.1-5.5) VFDs in favor of operative treatment.
Conclusions and Relevance
The findings of this randomized clinical trial suggest that operative treatment of patients with unstable chest wall injuries has modest benefit compared with nonoperative treatment. However, the potential advantage was primarily noted in the subgroup of patients who were ventilated at the time of randomization. No benefit to operative treatment was found in patients who were not ventilated.
Trial Registration
ClinicalTrials.gov Identifier: NCT01367951
Introduction
Unstable chest wall injuries (such as flail chest) result from blunt chest trauma and are associated with high rates of mortality and morbidity. These injuries have been shown to have a high incidence of complications, including chest wall instability, severe pulmonary restriction owing to paradoxical movement of the flail segment, loss of lung volume, and difficulties with pain management.1,2,3,4 The resulting combination of instability, decreased lung volume, and pain leads to decreased pulmonary function and can result in the need for prolonged ventilation. Prolonged mechanical ventilation is associated with high rates of pneumonia, sepsis, tracheostomy, barotrauma, protracted intensive care unit (ICU) stay, and high health care costs.1,2,3,4,5,6,7
The most prevalent treatment of severe chest wall injuries consists of nonoperative management via intubation and intermittent positive-pressure ventilation as needed, analgesia, pulmonary toilet, chest tube drainage as needed, and chest physiotherapy.2,4,8 However, such an approach may not produce optimal results, and there has been substantial interest in surgical fixation of rib fractures in the last decade to improve outcomes in patients with these injuries. Multiple retrospective studies and meta-analyses have reported improved outcomes with operative compared with nonoperative treatment, such as lower rates of pneumonia, sepsis, and duration of mechanical ventilation.1,5,9,10,11,12 However, randomized clinical trials (RCTs) in this area are limited and have reported somewhat conflicting results.3,6,7 Notably, there are several limitations to previous RCTs, including small sample sizes, single-center experience, outdated methods of surgical fixation, and lack of well-defined inclusion and exclusion criteria.
In order to address this, we conducted a multicenter, prospective, RCT comparing operative treatment of carefully selected, acute, unstable chest wall injuries with the current criterion standard of nonoperative management. We hypothesized that operative treatment would lead to reduced time receiving mechanical ventilation during the first 28 days after injury as compared with nonoperative treatment.
Methods
Trial Design
This was an international, multicenter, parallel, RCT comparing operative with nonoperative treatment in patients with unstable chest wall injuries. Trial protocol versions and statistical analysis plan are available in Supplement 1, Supplement 2, Supplement 3, Supplement 4, Supplement 5, Supplement 6, and Supplement 7. The study was approved by the institutional review board at all participating sites, and written informed consent was obtained for all study participants. This study followed the Consolidated Standards of Reporting Trials (CONSORT) reporting guidelines.
Participants
Patients were recruited from 15 level 1 trauma centers, from October 10, 2011, to September 17, 2018. Inclusion criteria were patients between the ages of 16 and 85 years meeting 1 of the following 2 indications for surgical fixation of their chest wall injury: (1) displaced rib fractures with a flail chest or (2) non–flail chest injuries with severe chest wall deformity. Patient data on race and ethnicity were not gathered as they were not anticipated to affect study outcomes. All injuries were assessed with computed tomography scan and had to meet clearly defined criteria to be eligible for the trial (eMethods in Supplement 8).
Exclusion criteria are detailed in the eMethods in Supplement 8. We excluded patients with rib fractures that were not amenable to fixation, patients with significant injuries who would otherwise require prolonged mechanical ventilation (severe pulmonary contusion, severe traumatic brain injury, upper airway injuries), patients with home oxygen requirements, and patients who were considered medically unfit for surgery. We excluded patients who could not be randomly assigned to an intervention group within 72 hours of injury and receive operative treatment within 96 hours of injury.
Patients were enrolled regardless of their need for mechanical ventilation at presentation. Surgeons performing surgical intervention were experienced in rib fracture fixation and had to meet prespecified criteria to be eligible to participate in the trial (eAppendix and eMethods in Supplement 8).
Randomization/Blinding
Randomization was performed by the research team via an online randomization system (http://www.randomize.net), stratified according to site (using a block size of 6 and blocking factor of 3) and concealed to investigators. Patients were randomly assigned to 1 of 2 treatment groups: (1) operative treatment and (2) nonoperative treatment. Given the nature of the trial, the patients and treating physicians were not blinded to the randomization allocation.
Interventions
Patients randomly assigned to the nonoperative group received standard of care treatment for unstable chest wall injuries, including mechanical ventilation as needed, pulmonary toilet or chest physiotherapy, chest tube drainage as needed, and pain management using a standardized analgesia protocol as a guide.
Patients randomly assigned to operative treatment received all of the treatments outlined previously as well as surgical stabilization of their chest wall injury with the use of a standardized surgical protocol. Only plates and screws were used for surgical fixation of fractures, with the goal of repairing a sufficient number of fractures to restore chest wall stability.
Attempts were made to standardize ventilation and weaning in both groups according to a specified protocol. Details of analgesia, treatment, and ventilation protocol guides are detailed in the eMethods and eFigures 1, 2, 3, and 4 in Supplement 8.
Outcomes
The primary outcome was the number of ventilator-free days (VFDs) during the first 28 days after injury. This measure combines survival and duration of ventilation and penalizes nonsurvivors.13 VFDs were defined as the number of days that a patient was alive and free from invasive ventilation in the first 28 days after injury. Patients who died before day 28, regardless of whether they achieved unassisted breathing, or patients who received invasive ventilation for more than 28 days, were assigned zero ventilator-free days. To be successfully extubated, patients had to remain extubated for at least 24 hours. Patients who were reintubated within 24 hours did not have that time counted toward a VFD. Only invasive mechanical ventilation was considered, and noninvasive modes such as continuous positive airway pressure or bilevel positive airway pressure were not included.
Secondary outcomes included total time on mechanical ventilation and length of stay in the ICU and hospital; incidence of acute pneumonia (within 28 days after injury), ventilator-associated pneumonia (VAP), sepsis, tracheostomy, and death; and surgical complications (wound infections, fixation failure, etc). Pneumonia, VAP, and sepsis were diagnosed using specific predefined criteria, as described in the eMethods in Supplement 8. Patients were evaluated daily until discharged from the ICU. They were then followed up at 2 weeks, 6 weeks, 3 months, 6 months, and 12 months after injury.
Sample Size
A pretrial power analysis was performed based on the length of time receiving mechanical ventilation (eMethods in Supplement 8). However, we noted that length of time receiving mechanical ventilation does not take into account patients who die early (with a short time receiving mechanical ventilation), which may paradoxically affect the results. Therefore, in 2015, the sample size was recalculated using blinded data from our pilot study of 24 patients, which revealed an SD of 5.08 for VFDs. Assuming this, a sample size of 103 per group (206 overall) was required to detect a difference of 2 VFDs, using a 2-tailed α error of .05 and a power of 0.80 between the 2 groups. A difference of 2 VFDs was considered the minimal clinically important difference to detect in this trial based on both a review of the critical care literature and formal survey of critical care physicians and trialists at the participating study centers. As this outcome is measured in-hospital, before discharge, loss to follow-up was not expected for the primary outcome, and therefore, no adjustment was made.
Statistical Analysis
Statistical analysis was performed from March 20, 2019, to March 5, 2021, after a modified intention-to-treat principle after the exclusion of patients who had withdrawn consent, undergone duplicate randomization owing to website system error, or had undergone randomization but were subsequently found to be ineligible. Baseline demographics and injury characteristics were summarized using descriptive statistics. The primary inferential between-group comparison of VFDs was performed using a Wilcoxon rank sum test owing to distributional concerns. This analysis was supplemented with a t test to accompany the estimate of treatment effect expressed as the difference in means between the groups with 95% CIs. This analysis framework was also consistent with the regression models used in the subgroup analyses. The primary outcome was available for all patients, and there were no missing values. Before unblinding, a prespecified subgroup analysis was performed to explore the effect of mechanical ventilation at time of randomization on the primary outcome. Therefore, a subgroup analysis by baseline intubation status was conducted by means of a linear regression model.
Secondary analyses were supportive and exploratory, using a t test, or a Wilcoxon rank sum test for continuous outcomes and χ2or Fisher exact tests for binary outcomes. A time-to-event analysis was used for length of hospital stay, by looking at time from admission to discharge from hospital. Given that some patients died before final discharge, death was a competing risk for discharge. To account for this, the cumulative incidence curve was used to estimate the time-to-discharge survival. Treatment groups were compared on time to discharge by means of a proportional hazard model, and the treatment effect was expressed as a hazard ratio (HR) with 95% CIs. For this outcome, a subgroup analysis by baseline intubation status was conducted using a proportional hazard model. Analyses were conducted using R statistical software, version 4.0.4 (R Foundation for Statistical Computing).
Results
A total of 211 patients were recruited from 15 sites across Canada and the US from 2011 to 2018. Recruitment was stopped once sample size was reached. A total of 111 patients were randomly assigned to operative treatment, and 100 patients were randomly assigned to nonoperative treatment (Figure). After exclusions, baseline and primary outcome data were obtained from 207 patients (operative group: 108 patients [52.2%]; mean [SD] age, 52.9 [13.5] years; 81 male [75%]; 27 female [25%]; nonoperative group: 99 patients [47.8%]; mean [SD] age, 53.2 [14.3] years; 75 male [76%]; 24 female [24%]).
Figure. Consolidated Standards of Reporting Trials (CONSORT) Flow Diagram.
The most common mechanisms of injury were motor vehicle collisions (operative, 40 [37%] vs nonoperative, 30 [30%]), falls (operative, 18 [17%] vs nonoperative, 26 [26%]), motorcycle collisions (operative, 14 [13%] vs nonoperative, 15 [15%]), and pedestrian injuries (operative, 12 [11%] vs nonoperative, 10 [10%]). Patients had sustained a mean (SD) of 10.3 (4.1) rib fractures. At the time of admission, patients in the operative group had a mean (SD) Injury Severity Score (ISS) of 25.3 (10.7) compared with 26.0 (10.9) in the nonoperative group. A total of 184 of 206 patients (89%) had a pneumothorax, 157 of 206 (76%) had a hemothorax, 111 of 206 (54%) had a pulmonary contusion, and 89 of 207 (43%) were receiving mechanical ventilation. There were no differences between the 2 groups in terms of baseline parameters or demographics (Table 1).
Table 1. Baseline Patient Characteristics.
| Patient characteristics | No. (%) | |
|---|---|---|
| Operative treatment (n = 108) | Nonoperative treatment (n = 99) | |
| Age, mean (SD), y | 52.9 (13.5) | 53.2 (14.3) |
| Sex | ||
| Male | 81 (75) | 75 (76) |
| Female | 27 (25) | 24 (24) |
| Current smoker | 37 (35) | 32 (33) |
| Diabetes | 19 (18) | 9 (9) |
| Prior chest injury | 18 (17) | 17 (17) |
| Prior lung disease | 18 (17) | 13 (13) |
| Mechanism of injury | ||
| Motor vehicle collision | 40 (37) | 30 (30) |
| Fall | 18 (17) | 26 (26) |
| Motorcycle | 14 (13) | 15 (15) |
| Pedestrian struck | 12 (11) | 10 (10) |
| Crush injury | 8 (7) | 4 (4) |
| Cycling | 6 (6) | 2 (2) |
| Other recreational activities | 0 (0) | 0 (0) |
| Assault | 0 (0) | 2 (2) |
| Pneumothorax | 93 (87) | 91 (92) |
| Hemothorax | 84 (79) | 73 (74) |
| Pulmonary contusion | 57 (53) | 54 (55) |
| Head injury | 7 (7) | 12 (12) |
| Chest tube | 105 (97) | 82 (83) |
| No. of rib fractures, mean (SD) | 10.1 (3.8) | 10.5 (4.3) |
| Injury Severity Score, mean (SD) | 25.3 (10.7) | 26.0 (10.9) |
| Glasgow Coma Scale, mean (SD) | 12.5 (3.9) | 12.3 (4.5) |
Patients in the operative group underwent surgical fixation of mean (SD) 4.3 (1.4) fractures. The 2 most common types of plates used for surgical fixation were pelvic reconstruction plates (57 of 108 [53%]) and precontoured locking rib plates (46 of 108 [43%]).
Mechanical Ventilation
At the time of randomization, 89 patients were intubated and receiving mechanical ventilation: 44 of 108 (41%) in the operative group and 45 of 99 (46%) in the nonoperative group. A further 31 patients of the remaining 118 patients (26%) who were not initially receiving mechanical ventilation required ventilator support during their hospitalization: 21 patients (19.4%) in the operative group and 10 patients (10.1%) in the nonoperative group.
Mean (SD) VFDs were 22.7 (7.5) days for the operative group and 20.6 (9.7) days for the nonoperative group, with a mean difference of 2.1 days between the 2 groups (95% CI, −0.3 to 4.5 days; P = .09). Mean (SD) total ventilator days were 5.6 (8.8) days for the operative group and 6.9 (10.4) days for the nonoperative group, with a mean difference of −1.3 days (95% CI −3.9 to 1.3 days; P = .34). There were no statistically significant differences between the 2 groups with regard to VFDs or number of days receiving mechanical ventilation (Table 2).
Table 2. Outcomes Related to Invasive Mechanical Ventilation.
| Outcomes | No. (%) | P value | ||
|---|---|---|---|---|
| Operative treatment (n = 108) | Nonoperative treatment (n = 99) | Mean difference (95% CI) | ||
| Need for mechanical ventilation | ||||
| Intubated at randomization | 44 (40.7) | 45 (45.5) | NA | NA |
| Required ventilation after randomizationa | 21 (19.4) | 10 (10.1) | NA | NA |
| Ever received mechanical ventilationa | 65 (60.2) | 55 (55.6) | NA | NA |
| Length of time on mechanical ventilation | ||||
| Ventilator-free days, mean (SD) | 22.7 (7.5) | 20.6 (9.7) | 2.1 (−0.3 to 4.5) | .09 |
| Ventilator days, mean (SD) | 5.6 (8.8) | 6.9 (10.4) | −1.3 (−3.9 to 1.3) | .34 |
Abbreviation: NA, not available.
Excluding mechanical ventilation needed for surgical procedures.
Length of Stay
Median (IQR) length of stay in the hospital was 16.5 (10.0-30.3) days for the operative group and 16.0 (8.5-32.5) days for the nonoperative group. Mean (SD) length of ICU stay was 8.4 (9.9) days in the operative group and 8.9 (10.4) days in the nonoperative group. In patients who were admitted to the ICU only, mean (SD) length of stay in the ICU was 10.8 (10.0) days in the operative group and 11.3 (10.5) days in the nonoperative group. There were no differences in length of hospital or ICU stay between groups (eResults and eTable 1 in Supplement 8).
Complications and Death
With the exception of tracheostomy (9 [8%] operative group vs 16 [16%] nonoperative group; 95% CI, 0.83-5.73; P = .13), the rates of complications were similar between the 2 groups (Table 3). There were 6 in-hospital deaths during the initial hospitalization: zero in the operative group and 6 (6.0%) in the nonoperative group (P = .01). Details are available in the eResults, eTable 2, and eTable 3 in Supplement 8. All but 1 of these patients was ventilated at the time of randomization.
Table 3. Complications.
| Complications | No. (%) | P value | ||
|---|---|---|---|---|
| Total (N = 207) | Operative treatment (n = 108) | Nonoperative treatment (n = 99) | ||
| Pneumoniaa | 10 (5) | 7 (6) | 3 (3) | .34 |
| VAP | 43 (21) | 23 (21) | 20 (20) | .98 |
| Sepsis | 24 (12) | 12 (11) | 12 (12) | .99 |
| Tracheostomy | 25 (12) | 9 (8) | 16 (16) | .13 |
| Empyema | 5 (2) | 2 (2) | 3 (3) | .67 |
| Deathb | 6 (3) | 0 (0) | 6 (6) | .01 |
Abbreviation: VAP, ventilator-associated pneumonia.
Within 28 days of injury.
In-hospital death from initial hospital admission.
Subgroup Analysis
A prespecified subgroup analysis was performed based on intubation and mechanical ventilation at the time of randomization. Among patients who were ventilated at the time of randomization, the mean difference in VFDs was 2.8 days (95% CI, 0.1-5.5 days), and among nonventilated patients, it was 0.6 days (95% CI, −1.8 to 2.9 days; P for subgroup interaction = .22), both in favor of the operative group (Table 4).
Table 4. Subgroup Analysis.
| Outcomes | Nonintubated at randomization | Intubated at randomization | P valuea | ||||
|---|---|---|---|---|---|---|---|
| Operative treatment (n = 64) | Nonoperative treatment (n = 54) | Mean difference (95% CI) | Operative treatment (n = 44) | Nonoperative treatment (n = 45) | Mean difference (95% CI) | ||
| Ventilator-free days, mean (SD) | 27.0 (2.3) | 26.4 (5.7) | 0.6 (−1.8 to 2.9) | 16.4 (8.0) | 13.6 (8.9) | 2.8 (0.1 to 5.5) | .22 |
| Ventilator days, mean (SD) | 1.0 (2.3) | 1.1 (4.4) | −0.1 (−2.8 to 2.6) | 12.4 (10.2) | 14.0 (11.1) | −1.6 (−4.7 to 1.5) | .47 |
| ICU length of stay, mean (SD), d | 3.4 (5.0) | 3.0 (6.0) | 0.4 (−2.5 to 3.3) | 15.6 (10.8) | 15.9 (10.3) | −0.3 (−3.7 to 3.0) | .75 |
| Hospital length of stay, median (IQR), d | 11.0 (9.0-17.0) | 11.0 (6.0-17.0) | 0.7b (95% CI, 0.5 to 1.0) | 30.0 (20.0-46.3) | 32.0 (16.0-57.0) | 1.4b (95% CI, 0.9 to 2.1) | .02 |
Abbreviation: ICU, intensive care unit.
P value for subgroup interaction.
Hazard ratio.
Among patients who were ventilated at the time of randomization, operative treatment reduced the length of hospital stay (increased the hazard of discharge), compared with nonoperative treatment (HR, 1.4; 95% CI, 0.9-2.1). The opposite effect was observed among patients who were not ventilated at randomization (HR, 0.7; 95% CI, 0.5-1.0; P for subgroup interaction = .02) (Table 4).
Between-group differences on the incidence of pneumonia, VAP, sepsis, and tracheostomy were similar in the subgroups of patients ventilated and nonventilated at the time of randomization. There was a trend toward higher mortality in the subgroup who were ventilated at the time of randomization (deaths: nonoperative treatment, 5 of 45 [11%] vs operative treatment, 0 of 44 [0%]; P = .06) (eTables 1, 2, and 3 in Supplement 8).
Surgical Complications and Reoperation
Surgical complications are detailed in the eResults in Supplement 8 but were generally infrequent. Four patients in total from the operative group required repeat surgery: 1 for irrigation and debridement of empyema, 1 for irrigation and debridement of empyema plus removal of loose hardware, 1 for video-assisted thoracic surgery evacuation of retained hemothorax, and 1 for loose hardware removal.
In the nonoperative group, 4 patients underwent an unplanned surgery: 2 for empyema evacuation, and 2 were treated with surgical fixation of their rib fractures owing to further deterioration despite maximal nonoperative management.
Discussion
This was, to our knowledge, the largest randomized trial to date of operative vs nonoperative treatment for acute unstable chest wall injuries. The findings of this trial suggest that operative treatment of patients with unstable chest wall injuries has modest benefit compared with nonoperative treatment. However, the potential advantage was primarily observed in the subgroup of patients who were receiving mechanical ventilation at the time of randomization.
A prespecified subgroup analysis revealed that mechanical ventilation at the time of randomization had a positive effect on certain outcomes. Nonventilated patients had no difference in any outcomes when comparing operative and nonoperative treatment. However, in patients who were receiving mechanical ventilation at the time of randomization, operative treatment was associated with higher VFDs (mean difference, 2.8 days; 95% CI, 0.1-5.5 days) and shorter hospital length of stay (HR, 1.4; 95% CI, 0.9-2.1), although the tests for subgroup interaction were negative for VFDs (P for subgroup interaction = .22), while being positive for hospital length of stay (P for subgroup interaction = .02).
There has been a substantially increased interest in surgical fixation of chest wall injuries and rib fractures recently, with a significant increase in the number of surgeries being performed.14 However, the literature supporting this shift in practice is limited and has numerous limitations. Four small randomized trials comparing operative with nonoperative treatment in flail chest injuries have been performed before our trial.3,6,7,15 Tanaka et al6 and Granetzny et al3 reported less time on mechanical ventilation, shorter ICU and hospital stay, and lower rates of pneumonia and tracheostomy with operative treatment compared with nonoperative treatment. Marasco et al7 reported no difference in their primary outcome of time on mechanical ventilation; however, surgery was associated with a shorter ICU length of stay and lower rates of tracheostomy. More recently, Liu et al15 reported shorter duration of mechanical ventilation and ICU stay and decreased rates of pneumonia with operative treatment. All 4 of these studies were performed at a single center with small numbers (50 patients or less), and 2 of the trials are more than 15 years old and employed outdated methods of surgical fixation.3,6 A meta-analysis of the 3 earlier published RCTs was conducted, with a total of 123 patients. The results favored operative over nonoperative treatment with respect to time receiving mechanical ventilation, ICU length of stay, hospital length of stay, and pneumonia.10 Of note, this meta-analysis has the limitations noted previously in the 3 RCTs it combines, and the total number of patients and centers is less than that in our trial reported here.
To our knowledge, our trial represents the largest prospective randomized study of operative vs nonoperative treatment for unstable chest wall injuries. The results of our trial are somewhat contradictory to the previous literature noted previously. Our trial did not show a statistically significant difference in VFDs (our primary outcome) in patients with unstable chest wall injuries with operative vs nonoperative treatment. Although there was a difference of 2.1 days between the 2 groups in favor of operative treatment, this did not reach statistical significance (95% CI, −0.3 to 4.5 days; P = .09). This may be attributable to the trial being underpowered as our sample size calculation was based on pilot data from 2 centers, whereas the trial was conducted across 15 centers, and the results we observed demonstrated higher variability than the initial pilot data.
In contrast to multiple previous small trials, we did not find convincing evidence for significant differences between the 2 groups regarding total time receiving mechanical ventilation, time in the ICU, total length of hospital stay, incidence of pneumonia, VAP, sepsis, empyema, or the need for tracheostomy. Our results did demonstrate a significant difference in mortality between the 2 groups: 6% (6 of 99) in the nonoperative group compared with zero (0 of 108) in the operative group (P = .01). However, it is difficult to identify the etiology for lowered mortality, given that there were no differences between the 2 groups regarding common causes of mortality in this patient population (eg, pneumonia, VAP, sepsis).
Although this trial did not demonstrate a statistically significant improvement in the primary outcome of VFDs with operative treatment, several important observations can be made. First, patients who require early mechanical ventilation had worse outcomes and higher complications compared with nonventilated patients, regardless of the treatment they received. Second, nonventilated patients did not demonstrate any benefit from operative treatment, as compared with nonoperative treatment, regarding any of the outcomes assessed. Third, ventilated patients seemed to benefit more from operative than nonoperative treatment in several aspects including improved VFDs (2.8 days) and a shorter hospital length of stay. Lastly, operative treatment was generally associated with a low rate of complications; in fact, the operative and nonoperative groups had an equivalent rate of reoperation (4 in each group). Taken together, we believe that these results suggest that the potential benefit to operative treatment may be in a select group of patients with an unstable chest wall injury who require early invasive mechanical ventilation. Routine surgical fixation of nonventilated patients is not supported given the results of this trial. Further investigation may be warranted with a larger sample size including only patients requiring mechanical ventilation.
Strengths and Limitations
Our study had several strengths. Strengths of our trial included the randomized prospective design, large number of patients (the largest sample size to date), multicenter nature, multidisciplinary input from a variety of subspecialities in study design, strict inclusion and exclusion criteria, modern and standardized operative treatment with plates and screws, and standardized protocols for patient care.
Our trial had several limitations. First, the sample size issues outlined previously are noteworthy, as we were underpowered to detect statistical significance in outcomes that were potentially clinically significant. The trial was also underpowered to detect significant differences in the subgroup of patients that were mechanically ventilated at the time of randomization. Second, the trial was conducted across a large number of centers, by variable types of surgeons (orthopedic, general, and thoracic surgery), and we are unable to comment on the effect of this. Although we attempted to standardize care across centers (standardized protocols for analgesia, operative treatment, ventilation, and weaning), it is likely that variations in the care of these patients occurred. However, this pragmatic element does improve the generalizability of our results. A further limitation was the inclusion of both ventilated and nonventilated patients, which led to comparison of a diverse cohort of patients. However, we recognized that trauma patients are intubated for a variety of reasons aside from their chest injury; we also wanted to capture patients who were not initially intubated but experienced early respiratory compromise and subsequently required mechanical ventilation. In addition, we were very interested in how surgery would affect these 2 groups differently, which is why we elected to include both groups but with a prespecified subgroup analysis. Finally, we were unable to blind assessors and those involved in patient care to treatment group, leading to the potential for bias in our primary outcome as the decision to wean or extubate a patient can be somewhat subjective, although steps were taken to mitigate this with a standardized weaning protocol.
Conclusions
In summary, this RCT represents, to our knowledge, the largest randomized trial to date of operative vs nonoperative treatment for acute unstable chest wall injuries. We found that operative treatment of unstable chest wall injuries provides modest benefit compared with nonoperative treatment. A potential advantage was observed with operative treatment in the subgroup of patients who were ventilated at the time of randomization. We found no benefit to operative treatment in patients who were not ventilated. We believe that these results support the potential use of surgery to stabilize rib fractures only in a select group of patients with acute unstable chest wall injuries who require mechanical ventilation. However, routine operative treatment of nonventilated patients is not indicated. Further studies in this area are warranted to determine the effects of surgery and to assess which patient populations who are ventilated are most likely to benefit from potential operative treatment.
Trial Protocol, Version 1
Trial Protocol, Version 2
Trial Protocol, Version 3
Trial Protocol, Version 4
Trial Protocol, Version 5
Trial Protocol, Version 6
Statistical Analysis Plan
eAppendix
eMethods
eFigure 1. Analgesia Protocol
eFigure 2. Critical-Care Pain Observation Tool
eFigure 3. Sedation Protocol
eFigure 4. Sedation-Agitation Scale
eResults
eTable 1. Length of Hospital and ICU Stay
eTable 2. Cause of In-hospital Death During Initial Hospital Admission
eTable 3. Complications Based on Intubation Status at the Time of Randomization
eReferences
Nonauthor Collaborators
Data Sharing Statement
References
- 1.Engel C, Krieg JC, Madey SM, Long WB, Bottlang M. Operative chest wall fixation with osteosynthesis plates. J Trauma. 2005;58(1):181-186. doi: 10.1097/01.TA.0000063612.25756.60 [DOI] [PubMed] [Google Scholar]
- 2.Nirula R, Diaz JJ Jr, Trunkey DD, Mayberry JC. Rib fracture repair: indications, technical issues, and future directions. World J Surg. 2009;33(1):14-22. doi: 10.1007/s00268-008-9770-y [DOI] [PubMed] [Google Scholar]
- 3.Granetzny A, Abd El-Aal M, Emam E, Shalaby A, Boseila A. Surgical vs conservative treatment of flail chest: evaluation of the pulmonary status. Interact Cardiovasc Thorac Surg. 2005;4(6):583-587. doi: 10.1510/icvts.2005.111807 [DOI] [PubMed] [Google Scholar]
- 4.Lafferty PM, Anavian J, Will RE, Cole PA. Operative treatment of chest wall injuries: indications, technique, and outcomes. J Bone Joint Surg Am. 2011;93(1):97-110. doi: 10.2106/JBJS.I.00696 [DOI] [PubMed] [Google Scholar]
- 5.Ahmed Z, Mohyuddin Z. Management of flail chest injury: internal fixation vs endotracheal intubation and ventilation. J Thorac Cardiovasc Surg. 1995;110(6):1676-1680. doi: 10.1016/S0022-5223(95)70030-7 [DOI] [PubMed] [Google Scholar]
- 6.Tanaka H, Yukioka T, Yamaguti Y, et al. Surgical stabilization of internal pneumatic stabilization? a prospective randomized study of management of severe flail chest patients. J Trauma. 2002;52(4):727-732. doi: 10.1097/00005373-200204000-00020 [DOI] [PubMed] [Google Scholar]
- 7.Marasco SF, Davies AR, Cooper J, et al. Prospective randomized controlled trial of operative rib fixation in traumatic flail chest. J Am Coll Surg. 2013;216(5):924-932. doi: 10.1016/j.jamcollsurg.2012.12.024 [DOI] [PubMed] [Google Scholar]
- 8.Simon B, Ebert J, Bokhari F, et al. ; Eastern Association for the Surgery of Trauma . Management of pulmonary contusion and flail chest: an Eastern Association for the Surgery of Trauma practice management guideline. J Trauma Acute Care Surg. 2012;73(5)(suppl 4):S351-S361. doi: 10.1097/TA.0b013e31827019fd [DOI] [PubMed] [Google Scholar]
- 9.Nirula R, Allen B, Layman R, Falimirski ME, Somberg LB. Rib fracture stabilization in patients sustaining blunt chest injury. Am Surg. 2006;72(4):307-309. doi: 10.1177/000313480607200405 [DOI] [PubMed] [Google Scholar]
- 10.Coughlin TA, Ng JWG, Rollins KE, Forward DP, Ollivere BJ. Management of rib fractures in traumatic flail chest: a meta-analysis of randomised controlled trials. Bone Joint J. 2016;98-B(8):1119-1125. doi: 10.1302/0301-620X.98B8.37282 [DOI] [PubMed] [Google Scholar]
- 11.Swart E, Laratta J, Slobogean G, Mehta S. Operative treatment of rib fractures in flail chest injuries: a meta-analysis and cost-effectiveness analysis. J Orthop Trauma. 2017;31(2):64-70. doi: 10.1097/BOT.0000000000000750 [DOI] [PubMed] [Google Scholar]
- 12.Slobogean GP, MacPherson CA, Sun T, Pelletier ME, Hameed SM. Surgical fixation vs nonoperative management of flail chest: a meta-analysis. J Am Coll Surg. 2013;216(2):302-11.e1. doi: 10.1016/j.jamcollsurg.2012.10.010 [DOI] [PubMed] [Google Scholar]
- 13.Yehya N, Harhay MO, Curley MAQ, Schoenfeld DA, Reeder RW. Reappraisal of ventilator-free days in critical care research. Am J Respir Crit Care Med. 2019;200(7):828-836. doi: 10.1164/rccm.201810-2050CP [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Dehghan N, Mah JM, Schemitsch EH, Nauth A, Vicente M, McKee MD. Operative stabilization of flail chest injuries reduces mortality to that of stable chest wall injuries. J Orthop Trauma. 2018;32(1):15-21. doi: 10.1097/BOT.0000000000000992 [DOI] [PubMed] [Google Scholar]
- 15.Liu T, Liu P, Chen J, Xie J, Yang F, Liao Y. A Randomized controlled trial of surgical rib fixation in polytrauma patients with flail chest. J Surg Res. 2019;242:223-230. doi: 10.1016/j.jss.2019.04.005 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Trial Protocol, Version 1
Trial Protocol, Version 2
Trial Protocol, Version 3
Trial Protocol, Version 4
Trial Protocol, Version 5
Trial Protocol, Version 6
Statistical Analysis Plan
eAppendix
eMethods
eFigure 1. Analgesia Protocol
eFigure 2. Critical-Care Pain Observation Tool
eFigure 3. Sedation Protocol
eFigure 4. Sedation-Agitation Scale
eResults
eTable 1. Length of Hospital and ICU Stay
eTable 2. Cause of In-hospital Death During Initial Hospital Admission
eTable 3. Complications Based on Intubation Status at the Time of Randomization
eReferences
Nonauthor Collaborators
Data Sharing Statement