Rib fractures are commonly encountered in the setting of thoracic trauma as a consequence of significant impacting forces on the chest wall. Rib fractures are found in 10% of all trauma patients and over 30% of chest trauma admissions.1 Concomitant injuries are often identified and complications range from minor discomfort and pain to pneumothorax, pneumonia, and death.2 The overall mortality associated with rib fractures is 10% and increases for each additional rib fracture.1,3 Early intervention and effective pain management are shown to be associated with fewer days of mechanical ventilation and cost-saving of at least US $10,000 per patient.4–7
Neuraxial and regional techniques primarily aim to minimize pain to improve outcomes. Inadequate pain control is associated with atelectasis, decreased mobility, and secretion clearance leading to impaired pulmonary mechanics. This lethal cascade leads to more serious complications including development of pneumonia, intubation, and increased mortality. In fact, pain control is the cornerstone of rib fracture management.8 The choice of analgesia is dependent on the physician’s expertise, injury complex, risk of complications, and ease of performing the procedure.
Epidural analgesia (EA) is shown to be an effective pain control intervention for rib fractures.9 Oftentimes, trauma patients are not candidates for EA secondary to concomitant injuries including spinal cord injuries, thoracic vertebral fractures, unstable pelvis, severe head injury, and coagulopathy.10 Paravertebral block (PVB) does not carry all of these contraindications, yet no comprehensive study has shown whether PVB bestows similar analgesic and outcome benefits in trauma patients.11,12 In elective thoracotomies for lung resections, EA and PVB have been found to provide similar pain control and outcomes.10 However, procedural complications including holding analgesia secondary to hypotension were found to be higher in patients receiving EA.13,14
Current clinical literature comparing the effect of EA and PVB on the clinical outcomes of rib fractures is scarce. In patients with significant bilateral chest wall trauma, EA seems most appropriate clinically. However, in patients with unilateral chest wall trauma, PVB is often utilized. There is minimal trauma literature to support this practice and it is unclear whether these 2 procedures provide similar outcomes. Utilization of the National Trauma Data Bank (NTDB) will not allow a comparison of procedural complications, including stopping analgesia for associated hypotension, but outcomes can be compared. The purpose of this study was to assess the association between the 2 procedures (PVB and EA) and hospital mortality as the primary outcome and length of stay (LOS), intensive care unit (ICU) admission, ICU LOS, mechanical ventilation, duration of mechanical ventilation, development of pneumonia, and development of other complications as secondary outcomes among adult trauma patients registered in the NTDB.
The NTDB is the largest prospectively held repository of trauma patients housing data contributed from over 900 trauma centers. Following the approval of the Institutional Review Board of Geisinger Medical Center, we queried the 2011 and 2012 versions of NTDB for all available records (Committee on Trauma, American College of Surgeons. NTDB Version RDS 2011. Chicago, IL, 2012; and Committee on Trauma, American College of Surgeons. NTDB Version RDS 2012. Chicago, IL, 2013). We included all patients above 18 years of age with an International Classification of Diseases, Ninth Revision (ICD-9) code of 807 to include rib fractures. Patients with sternum, larynx, and trachea fractures were excluded using ICD-9 807.2, 807.3, 807.5, and 807.6 codes. We also excluded patients with missing data. Finally, EA and PVBs were identified using ICD-9 procedure codes 03.91 and 04.81, respectively.
The extracted data from the NTDB included demographics, mechanism of injury (fall, motorcycle collision, motor vehicle collision), Glasgow Coma Scale, Injury Severity Score, Abbreviated Injury Scale, and premorbid conditions. We also gathered information on the presence of pulmonary contusion, hemothorax, pneumothorax, chest tube, flail chest, number of rib fractures, open rib fractures, and rib fixation.
The primary outcome of interest was in-hospital mortality. Secondary outcomes of interest were LOS, ICU admission, ICU LOS, mechanical ventilation, duration of mechanical ventilation, development of pneumonia, and development of any other complication.
The goal of propensity score matching is to gain marginal balance on available potential confounding variables.15 The propensity score model included all variables shown in Figure 1. Two propensity score models were created. The first model estimated the probability of receiving a PVB. Once propensity scores were obtained, each PVB patient was matched to one EA patient using a 1:1 greedy match algorithm where the maximum allowable difference in propensity scores for a valid match was 0.10.
EA and PVB patients in the matched data set were then grouped together as patients who had a procedure. A second propensity score model was created that estimated the probability of having a procedure. Again, propensity scores were obtained, and each “procedure” patient was matched to one “nonprocedure” patient by using a 1:1 greedy match algorithm where the maximum allowable difference in propensity scores for a valid match was 0.10. Standardized mean differences were calculated before and after matching to assess the balance on baseline variables. The mean distance in propensity scores was also assessed for the EA/PVB group (mean: 0.0077 ± 0.0210) and the procedure/no-procedure (Proc/NoProc) group (mean: 2.3 × 10–6 ± 4.6 × 10–5). Propensity score matching was considered successful if all matched values had a standardized mean difference <0.10.
The primary outcome of interest was in-hospital mortality. Mortality was assessed using χ2 tests in the 2 matched data sets. Secondary outcomes of interest were LOS, ICU admission, ICU LOS, mechanical ventilation, duration of mechanical ventilation, development of pneumonia, and development of any other complication. Nonnormalized outcomes, such as LOS, ICU LOS, duration of mechanical ventilation, were analyzed using Mann-Whitney U test. Categorical outcomes such as ICU admission, mechanical ventilation, development of pneumonia, and development of any other complication were analyzed using χ2 tests. These outcomes were compared between EA/PVB and also Proc/NoProc. The primary outcome was considered significant with a P value of <.05. Bonferroni correction was used to maintain an overall α-level of .05 for the secondary outcomes. These outcomes were considered statistically significant where P < .007.
Nonnormal data were expressed as median with interquartile range (IQR) and ratio of geometric means. Categorical data were expressed as frequencies and percentages as well as odds ratios (ORs). Nonnormal data were analyzed using Mann-Whitney U test, and χ2 test was used for categorical variables. Multivariable logistic regression was used to test the association of procedure and mortality in the unmatched data set. No power calculation was done in planning this study, and all available records from the NTDB were included. The observed mortality for the data set was 5%, which was then used in our power calculation. Setting α = .05 and 1 − β = .90, our study was able to detect an OR of 2.1 in the PM EA/PVB group and an OR of 1.75 in the PM Proc/No Proc group. SAS 9.4 (SAS Institute, Cary, NC) was used for all the statistical analyses.
The study flow diagram is shown in Figure 2. There were a total number of 1,617,999 records in the 2011 and 2012 versions of NTDB. After stratifications and exclusions, 1110 patients were placed in the PVB group, 1073 patients in the EA group, and 192,583 remained in the no-procedure group.
Tables 1 and 2 summarize the baseline characteristics and the overall outcomes. The majority of patients in each group were men (>60%) over the age of 50 years who were admitted after a motor vehicle collision (>30%). Pulmonary contusion, pneumothorax, and hemothorax were less common in the no-procedure group. Tube thoracotomy and rib fixation were more frequently performed in the EA and PVB groups. More than 60% of patients who had a procedure (EA or PVB) were admitted to ICU, whereas almost 46% of patients in the no-procedure group were admitted to ICU. Nearly the same proportion of patients in each group received mechanical ventilation (22.6%, 24.6%, and 21.7% in EA, PVB, and no-procedure groups, respectively). Mortality was 5.1% in the no-procedure group compared to 2.1% in both EA and PVB groups.
Following the propensity score matching between the EA and PVB patients, 557 patients remained in each group. Comparison of primary and secondary outcomes between the propensity-matched patients is summarized in Table 3. There was no association found for the primary outcome of mortality (OR 1.51; 95% confidence interval, 0.61–3.73; P = .37). There was also significant association found with any of the secondary outcomes.
In the next step, previously propensity score–matched EA and PVB patients were combined into the procedure group comprising 1114 patients. We ran another propensity score matching to find 1114 patients from the no-procedure group. Table 4 summarizes the comparison of the newly propensity score-matched patients. Receiving a procedure was associated with a significantly prolonged LOS and frequency of ICU admission (both P < .0001) but not the days spent in ICU (4 days, IQR: 2–8 days vs 4 days, IQR: 2–9 days; P = .002). Having no procedure was associated with a significantly higher mortality (OR: 2.25; 95% confidence interval, 1.14–3.84; P = .002).
Finally, we tested for the association between having no procedure and mortality using different methods. These data are summarized in Table 5. Odds of mortality with no procedure was significantly increased across all calculation methods (all P < .01).
In this NTDB study, we found that EA and PVB were both associated with improved survival following rib fractures compared to having no procedure. EA and PVB were associated with similar clinical outcomes. In the setting of rib fracture, inadequate pain control leads to reduced tidal volume and secretion clearance to minimize chest wall motion. Therefore, adequate pain control is critical to avoid life-threatening complications, most notably respiratory failure, pneumonia, and death. Options for analgesia include EA, PVB, intercostal nerve block, intrapleural infusion, and IV narcotics.12,16–19
Mohta et al12 performed a randomized pilot study on 30 adults with 3 or more unilateral rib fractures and reported comparable outcomes in pain control and higher incidence of hypotension with EA. No other randomized study has been conducted since to compare these methods. A study was done by Dehghan et al20 on flail chest injuries using NTDB. Although the authors had not used propensity score matching in their study and had used only a subset of extreme cases, ie, flail chest compared to our study, ie, rib fractures, they also found that only about 8% of the patients had received epidural catheters for pain management.20
EA is shown to be associated with shorter duration of mechanical ventilation and a lower incidence of pneumonia.9 Early studies were in favor of EA over PVB,21 yet highlighting the need for close monitoring with EA especially in the elderly.16 A systematic review by Carrier et al16 showed the only benefit of EA over other procedures was shorter duration of mechanical ventilation, yet this finding was offset by a significantly higher rate of hypotension. Later studies showed that PVB has same or better outcomes compared to EA.10,22 This is also supported by a meta-analysis and a review of best evidence in elective thoracic surgeries.10,14 A recent retrospective study by Zaw et al23 reported that patients with epidural catheters had longer hospitalization yet reduced mortality. They reported higher DVT rates, which can possibly be due to extension of the study to 2004.23 Over the past decade, rigorous DVT prophylaxis measures have been adopted by most medical institutes.
It has been shown that correct administration of PVBs can achieve desirable unilateral analgesia for different purposes including pain management of rib fractures.10,11,24 One advantage of PVB (either single injection or continuous catheter) over EA is that it provides adequate analgesia with a reduced risk of hypotension and bradycardia.25,26 Absolute contraindications of PVB are patient refusal, local infection, allergic reactions to local anesthetics, and tumors in the paravertebral space, which are fewer in comparison to previously described contraindications of EA.10 PVB fell out of favor because of questioning of its adequacy of pain control,27 yet later studies supported reaching the desired analgesic level with PVB.11,12 Currently, Procedure-Specific Postoperative Pain Management working group has grade A evidence to recommend both PVB and EA to achieve adequate pain control.28
Although we found that EA and PVB were associated with similar clinical outcomes, NTDB does not capture procedural complications including associated hypotension requiring holding or stopping analgesia. Elective thoracic literature has found that EA and PVB provide similar postoperative pain control in small randomized trials.10,14 Randomized trials in the setting of trauma are needed to determine these procedural complications because it is possible that there are certain patient populations (eg, geriatric patients) that may benefit from PVB over EA.
We found that having no procedure was associated with a significantly higher odds of mortality. Factors that are known to increase mortality of rib fractures are age,29 premorbid conditions,30 number of rib fractures,31 flail chest,32 and development of pneumonia.29 Since the comparison in our study was performed on propensity score–matched patients, we had already controlled for the likely effect of these conditions on mortality. In fact, our propensity score matching included all the available variables that were retrieved from NTDB, which goes beyond the aforementioned known factors. Our study also highlights the need to nationally improve our pain management practice of rib fractures since nearly 99% of patients with rib fractures in the NTDB received neither EA nor PVB. Although the underutilization of PVB can be partially explained by being an intermediate-high skill level block,33 we are still far behind the optimal prevalence of EA or PVB in management of rib fractures.
Although NTDB is the largest available repository of trauma patients, studies using NTDB carry a number of limitations. NTDB is known to be a convenience sample of larger hospitals with disproportionately more severely injured patients; therefore, selection bias is always part of any NTDB study. In addition, there is the possibility that the missing data in the NTDB are not random.34 There is also the possibility that the observed association between having no procedure and mortality is due to unrecorded or unobserved confounders. In addition, survivors of the initial trauma are likely to die within the first 48 hours of hospital admission, which cannot be picked using the NTDB.35 These limitations can be overcome through well-designed prospective trials comparing the effects of EA, PVB, and no procedures.
The study of 2011 and 2012 NTDB records showed that less than 1.2% of patients with rib fractures received EA or PVB for pain control. EA and PVB were found to be associated with significantly better survivals in patients who received either procedure, yet, this might be due to selection of healthier patients or unavailability of these techniques in certain centers. Therefore, prospective studies are required to confirm these observed associations.
Name: Mahdi Malekpour, MD.
Contribution:This author helped withacquisition of data, interpretation, drafting of manuscript, and revision.
Name:Ammar Hashmi, MD.
Contribution:This author helped withacquisition of data, interpretation, and drafting of manuscript.
Name:James Dove, BA.
Contribution:This author helpedwith analysis and interpretation of data.
Name:Denise Torres, MD, FACS.
Contribution:This author helpedwith design of study, interpretation of data, and drafting of manuscript.
Name:Jeffrey Wild, MD, FACS.
Contribution:This author helpedwith design of study, interpretation of data, drafting of manuscript, and revision.
This manuscript was handled by:Richard P. Dutton, MD.
1. Ziegler DW, Agarwal NN. The morbidity and mortality of rib fractures. J Trauma. 1994;37:975–979.
2. Unsworth A, Curtis K, Asha SE. Treatments for blunt chest trauma and their impact on patient outcomes and health service delivery. Scand J Trauma Resusc Emerg Med. 2015;23:17.
3. Flagel BT, Luchette FA, Reed RL, et al. Half-a-dozen ribs: the breakpoint for mortality. Surgery. 2005;138:717–723
4. 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;52:727–732.
5. 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;216:924–932.
6. Majercik S, Wilson E, Gardner S, Granger S, VanBoerum DH, White TW. In-hospital outcomes and costs of surgical stabilization versus nonoperative management of severe rib fractures. J Trauma Acute Care Surg. 2015;79:533–538.
7. Wada T, Yasunaga H, Inokuchi R, et al. Effectiveness of surgical rib fixation on prolonged mechanical ventilation in patients with traumatic rib fractures: a propensity score-matched analysis. J Crit Care. 2015;30:1227–1231.
8. Karmakar MK, Ho AM. Acute pain management of patients with multiple fractured ribs. J Trauma. 2003;54:615–625.
9. Bulger EM, Edwards T, Klotz P, Jurkovich GJ. Epidural analgesia improves outcome after multiple rib fractures. Surgery. 2004;136:426–430.
10. Davies RG, Myles PS, Graham JM. A comparison of the analgesic efficacy and side-effects of paravertebral vs epidural blockade for thoracotomy–a systematic review and meta-analysis of randomized trials. Br J Anaesth. 2006;96:418–426.
11. Karmakar MK, Critchley LA, Ho AM, Gin T, Lee TW, Yim AP. Continuous thoracic paravertebral infusion of bupivacaine for pain management in patients with multiple fractured ribs. Chest. 2003;123:424–431.
12. Mohta M, Verma P, Saxena AK, Sethi AK, Tyagi A, Girotra G. Prospective, randomized comparison of continuous thoracic epidural and thoracic paravertebral infusion in patients with unilateral multiple fractured ribs–a pilot study. J Trauma. 2009;66:1096–1101.
13. Moen V, Dahlgren N, Irestedt L. Severe neurological complications after central neuraxial blockades in Sweden 1990-1999. Anesthesiology. 2004;101:950–959.
14. Scarci M, Joshi A, Attia R. In patients undergoing thoracic surgery is paravertebral block as effective as epidural analgesia for pain management? Interact Cardiovasc Thorac Surg. 2010;10:92–96.
15. D’Agostino RB Jr, D’Agostino RB Sr.. Estimating treatment effects using observational data. JAMA. 2007;297:314–316.
16. Carrier FM, Turgeon AF, Nicole PC, et al. Effect of epidural analgesia in patients with traumatic rib fractures: a systematic review and meta-analysis of randomized controlled trials. Can J Anaesth. 2009;56:230–242.
17. Truitt MS, Murry J, Amos J, et al. Continuous intercostal nerve blockade for rib fractures: ready for primetime? J Trauma. 2011;71:1548–1552.
18. Short K, Scheeres D, Mlakar J, Dean R. Evaluation of intrapleural analgesia in the management of blunt traumatic chest wall pain: a clinical trial. Am Surg. 1996;62:488–493.
19. Yang Y, Young JB, Schermer CR, Utter GH. Use of ketorolac is associated with decreased pneumonia following rib fractures. Am J Surg. 2014;207:566–572.
20. Dehghan N, de Mestral C, McKee MD, Schemitsch EH, Nathens A. Flail chest injuries: a review of outcomes and treatment practices from the National Trauma Data Bank. J Trauma Acute Care Surg. 2014;76:462–468.
21. Marret E, Bazelly B, Taylor G, et al. Paravertebral block with ropivacaine 0.5% versus systemic analgesia for pain relief after thoracotomy. Ann Thorac Surg. 2005;79:2109–2113.
22. Conlon NP, Shaw AD, Grichnik KP. Postthoracotomy paravertebral analgesia: will it replace epidural analgesia? Anesthesiol Clin. 2008;26:369–380, viii.
23. Zaw AA, Murry J, Hoang D, et al. Epidural analgesia after rib fractures. Am Surg. 2015;81:950–954.
24. Elsayed H, McKevith J, McShane J, Scawn N. Thoracic epidural or paravertebral catheter for analgesia after lung resection: is the outcome different? J Cardiothorac Vasc Anesth. 2012;26:78–82.
25. Cheema SP, Ilsley D, Richardson J, Sabanathan S. A thermographic study of paravertebral analgesia. An esthesia. 1995;50:118–121.
26. Saito T, Den S, Cheema SP, et al. A single-injection, multi-segmental paravertebral block-extension of somatosensory and sympathetic block in volunteers. Acta Anesthesiol Scand. 2001;45:30–33.
27. Richardson J, Lönnqvist PA. Thoracic paravertebral block. Br J Anaesth. 1998;81:230–238.
28. Kehlet H, Wilkinson RC, Fischer HB, Camu F; Prospect Working Group. PROSPECT: evidence-based, procedure-specific postoperative pain management. Best Pract Res Clin Anesthesiol. 2007;21:149–159.
29. Battle CE, Hutchings H, Evans PA. Risk factors that predict mortality in patients with blunt chest wall trauma: a systematic review and meta-analysis. Injury. 2012;43:8–17.
30. Brasel KJ, Guse CE, Layde P, Weigelt JA. Rib fractures: relationship with pneumonia and mortality. Crit Care Med. 2006;34:1642–1646.
31. Holcomb JB, McMullin NR, Kozar RA, Lygas MH, Moore FA. Morbidity from rib fractures increases after age 45. J Am Coll Surg. 2003;196:549–555.
32. Liman ST, Kuzucu A, Tastepe AI, Ulasan GN, Topcu S. Chest injury due to blunt trauma. Eur J Cardiothorac Surg. 2003;23:374–378.
33. De Cosmo G, Aceto P, Gualtieri E, Congedo E. Analgesia in thoracic surgery: review. Minerva Anestesiol. 2009;75:393–400.
34. Roudsari B, Field C, Caetano R. Clustered and missing data in the US National Trauma Data Bank: implications for analysis. Inj Prev. 2008;14:96–100.
35. Evans JA, van Wessem KJ, McDougall D, Lee KA, Lyons T, Balogh ZJ. Epidemiology of traumatic deaths: comprehensive population-based assessment. World J Surg. 2010;34:158–163.