I. STATEMENT OF THE PROBLEM AND QUESTIONS TO BE ADDRESSED
Studies of the consequences and treatment of blunt thoracic trauma (BTT) remain hampered by a varying pathologic definition of the disease. Entities typically classified as BTT include chest wall lesions such as rib fractures, flail chest and soft-tissue contusion; intrapleural lesions such as hemothorax and pneumothorax; parenchymal lung injuries such as pulmonary contusion and lung laceration; and mediastinal lesions such as blunt cardiac injury.1,2 For purposes of this evidence-based review, we are concerned primarily with those injuries to the chest wall that produce their morbidity through pain and its associated mechanical ventilatory impairment. Thus, blunt chest trauma is defined here to include soft-tissue trauma and injuries to the bony thorax such as rib fractures and flail chest.3
Within the scope of this definition, the incidence and morbidity of BTT clearly remain significant. Rib fractures themselves are believed to be very common and have been documented in up to two thirds of cases of chest trauma.4,5 In another review, 10% of all patients admitted to one trauma center had radiographic demonstration of rib fractures.3 Isolated single or multiple rib fractures are one of the most common injuries in the elderly, at approximately 12% of all fractures, with an increasing incidence recorded as the population ages.6 The true incidence of bony thoracic injury may be underreported, as up to 50% of fractures may be undetected radiographically.7
For patients with blunt chest wall trauma, the morbidity and mortality are significant. These injuries are associated with pulmonary complications in more than one third of cases3 and pneumonia in as many as 30% of cases.3,8,9 Patients older than 65 years may be even more prone to major complications after blunt chest wall injury,3,10–12 with 38% respiratory morbidity from isolated rib fractures in another review.13 Because blunt chest wall trauma causes death indirectly, through pulmonary and nonpulmonary complications, the true mortality rate for these injuries is hard to evaluate. In one study, 6% of patients with blunt chest trauma died, and at least 54% of these deaths could be directly attributed to secondary pulmonary complications.3 An elderly group of patients suffered an 8% mortality rate from isolated rib fractures.13 Mortality of isolated flail chest has been as high as 16%.14 The incremental costs attached to pulmonary complications of blunt chest trauma have not been addressed in the literature but clearly would be measured in “intensive care unit (ICU) days” and “ventilator days,” both of which are expensive commodities.
The treatment for injuries of the bony thorax has varied over the years, ranging from various forms of mechanical stabilization15,16 to obligatory ventilatory support.17–19 It is now generally recognized that pain control, chest physiotherapy, and mobilization are the preferred mode of management for BTT.9,20 Failure of this regimen and ensuing mechanical ventilation sets the stage for progressive respiratory morbidity and mortality.3,8,20 Consequently, several different strategies of pain control have been used, including intravenous narcotics, local rib blocks, pleural infusion catheters, paravertebral blocks, and epidural analgesia. Each of these modalities has its own unique advantages and disadvantages, and the overall most efficacious method has not previously been clearly identified. Subsequently, analgesic practices vary widely in this crucial setting. In one recent review, the majority of BTT patients were still managed with intravenous or oral narcotics.21 Other authors noted that epidural catheters were offered in only 22% of elderly BTT patients and 15% of a younger cohort.9
This review seeks to identify the optimal method(s) of pain control for patients with blunt chest trauma. The specific questions that are addressed using an evidence-based approach for outcome evaluation are as follows:
- Which patients with blunt chest trauma are at particular risk for respiratory morbidity caused by pain and deserve special attention to pain management?
- With consideration for safety, feasibility, and therapeutic effectiveness, what is the optimal method of pain control in blunt chest trauma?
- For the recommended modality/modalities, what technical recommendations can be made for the administration of analgesia in blunt chest trauma? A. Anesthetic and technology concerns. B. Nursing considerations.
A computerized search was conducted of the MEDLINE, EMBASE, and Cochrane Controlled Trials databases for North American and European English language literature for the period from 1966 through December 31, 2004. The initial search terms were “chest injuries,” “thoracic injuries,” “rib fractures,” and “flail chest.” These were cross-referenced for the secondary terms “analgesia,” “anesthesia,” and “pain.” This search initially yielded 213 articles. One hundred twenty-eight of these articles were excluded as being case studies, reviews, letters, or otherwise irrelevant to the questions being asked. This yielded a file of 85 articles for review. An additional 52 articles were obtained from the references of these studies, yielding a total of 137 studies for review and grading. Ninety-five of these were deemed appropriate for inclusion in the final evidentiary tables.
The practice parameter workgroup for analgesia in blunt thoracic trauma consisted of five trauma surgeons, one trained as a thoracic surgeon, two anesthesiologists, and one trauma clinical nurse specialist. All studies were reviewed by two committee members and graded according to the standards recommended by the EAST Ad Hoc Committee for Guideline Development.22 Grade I evidence was also subgraded for quality of design using the Jahad Validity Scale published in Controlled Clinical Trials in 1996.23 Any studies with conflicting grading were reviewed by the committee chairperson and were all Grade I studies. Recommendations were formulated based on a committee consensus regarding the preponderance and quality of evidence.
A. Efficacy of Analgesic Modalities
- Use of epidural analgesia (EA) for pain control after severe blunt injury and nontraumatic surgical thoracic pain significantly improves subjective pain perception and critical pulmonary function tests compared with intravenous narcotics. EA is associated with less respiratory depression, somnolence, and gastrointestinal symptoms than intravenous narcotics. EA is safe, with permanent disability being extremely rare and negligible mortality attributable to treatment.
- Epidural analgesia may improve outcome as measured by ventilator days, ICU length of stay, and hospital length of stay.
- There is some Class I and adequate Class II evidence to indicate that paravertebral or extrapleural infusions are effective in improving subjective pain perception and may improve pulmonary function.
- Although paravertebral or extrapleural analgesia is effective, there is an inadequate quantity of comparative evidence or information regarding safety to establish any recommendation with regard to overall efficacy.
- The information regarding both the effectiveness and safety of intrapleural and intercostal analgesia is contradictory, and experience with trauma patients is minimal. Consequently, no recommendation can be made regarding overall efficacy of this modality.
B. Clinical Application of Pain Management Modalities to Treatment of Blunt Thoracic Trauma
- Epidural analgesia is the optimal modality of pain relief for blunt chest wall injury and is the preferred technique after severe blunt thoracic trauma.
- Patients with four or more rib fractures who are ≥ 65 years of age should be provided with epidural analgesia unless this treatment is contraindicated.
- Younger patients with four or more rib fractures or patients aged ≥ 65 years with lesser injuries should also be considered for epidural analgesia.
- The approach for pain management in BTT requires individualization for each patient. Clinical performance measures (pain scale, pulmonary examination/function, arterial blood gases) should be measured as judged appropriate at regular intervals.
- Presence in elderly patients of cardiopulmonary disease or diabetes should provide additional impetus for epidural analgesia, as these comorbidities may increase mortality once respiratory complications have occurred.
- Intravenous narcotics, by divided doses or demand modalities, may be used as initial management for lower risk patients presenting with stable and adequate pulmonary performance, provided the desired clinical response is achieved.
- High-risk patients who are not candidates for epidural analgesia should be considered for paravertebral (extrapleural) analgesia commensurate with institutional experience.
- A specific recommendation cannot be made for intrapleural or intercostal analgesia based on the available evidence, but its apparent safety and efficacy in the setting of thoracic trauma has been reported.
C. Technical Aspects of Epidural Analgesic Agents
There is insufficient Class I and Class II evidence to establish any specific techniques of epidural analgesia as a standard of care.
- Combinations of a narcotic (i.e., fentanyl) and a local anesthetic (i.e., bupivacaine) provide the most effective epidural analgesia and are the preferred drug combinations for use by this route. Use of such combinations allows decreased doses of each agent and may decrease the incidence of side effects attributable to each.
- Nursing care of the patient with an epidural catheter should involve frequent monitoring of appropriate parameters at intervals based on institutional judgment. These parameters should include, but may not be limited to, respiratory function, sedation level and urinary retention for epidural narcotics, and fluid balance and motor strength for epidural anesthetics. Lower extremity weakness may be seen with excessive sympathetic block but may also indicate epidural hematoma or abscess and should prompt appropriate evaluation. It should be noted that epidural anesthetics may mask the sensory deficits caused by these mass lesions such that monitoring of motor function is especially important when such agents are used. Vital signs should be monitored frequently, as hypotension may occur as a result of sympathetic block and fever may indicate catheter site infection.
- The duration of epidural therapy should be individualized for each patient on the basis of the clinical situation, response to therapy, and anticipated therapeutic and adverse response to systemic alternatives. Epidural analgesia should be discontinued and the catheter removed when it no longer offers a benefit over systemic medications.
- The epidural catheter should be removed and the tip sent for culture if the site becomes erythematous or indurated or if there are signs of systemic infection such as fever, rigors, or leukocytosis. The indications for antibiotic therapy in possible epidural catheter/epidural space infections are beyond the scope of this guideline and additional resources should be referenced.
- Although reliable literature describes the safe use of epidural analgesia on regular surgical floors, most victims of blunt thoracic trauma receiving this modality of treatment will have other primary indications for a higher level of care. Consequently, such patients in general should be nursed in a monitored setting with cardiac monitoring and continuous pulse oximetry.
- There is insufficient evidence at this time to make a recommendation regarding the use of continuous epidural infusion versus intermittent injection in trauma patients.
IV. SCIENTIFIC FOUNDATION
A. Historical Perspective
The treatment of blunt thoracic trauma has undergone dramatic evolution over the twentieth century. In the first half of the century, the primary emphasis was on mechanical stabilization of the bony injury. This was first done by such external devices as sandbags or traction systems and later by various surgical methods such as wires or screws.24 After 1950, the concept of “internal pneumatic stabilization” with positive-pressure mechanical ventilation was developed.25 This became more prevalent and obligatory mechanical ventilation became the standard for chest wall trauma.26
The management of severe, blunt thoracic trauma evolved into the modern era with the publication of two studies in 1975. In a small series, Trinkle27 demonstrated that optimal pain control, chest physiotherapy, and noninvasive positive-pressure ventilation could avert the need for intubation and mechanical ventilation. Also in 1975, Dittman28 published the first in a series of three articles on pain management in blunt chest trauma. In the first study, 19 patients with multiple rib fractures and flail segments were treated with continuous epidural analgesia and intubation and mechanical ventilation were withheld. Using objective clinical criteria to monitor progress (e.g., vital capacity, respiratory rate, and tidal volume), 17 patients were successfully managed without positive-pressure ventilation. Dittman29 subsequently showed that 46 of 49 (94%) spontaneously breathing patients maintained a vital capacity greater than 13 mL/kg and avoided positive-pressure ventilation through the use of morphine analgesia by means of a thoracic epidural catheter. Other European studies demonstrated good clinical results with epidural analgesia in blunt chest wall injuries when combined with pulmonary toilet and selective mechanical ventilation.30–32
Thus, the management of blunt thoracic trauma today focuses on both the underlying lung injury and on optimization of mechanics through chest physiotherapy and optimal analgesia.30,33–36 The critical importance of measuring ventilatory function tests as an objective means of monitoring adequacy of this analgesia was emphasized by the authors of the early studies.33–36 Subsequent studies of pain management in blunt thoracic trauma patients would use the same methodology and additionally focus on comparisons between modalities and on objective outcome parameters.37–40
B. Modalities of Analgesia
Intravenous narcotics have historically been the initial and most prevalent modality for relief of surgical and traumatic pain of all types. They are administered either by intermittent injection when pain is noted by the patient41 or continuous infusion.42 Most recently intravenous patient-controlled analgesia (PCA) has been developed to exploit the benefits of both methods.43,44 In this modality, a baseline intravenous infusion of morphine is provided and the patient may elicit an additional bolus for breakthrough pain.
The obvious advantages of intravenous narcotics are ease of administration and monitoring by nursing without the risks of an invasive procedure or need for specialized personnel. The efficacy of this modality for blunt chest wall trauma is controversial. Intravenous narcotics have been shown to improve pain scores and vital capacity, yet some clinicians consider them inadequate in this setting.41,43 The disadvantages of systemic narcotics are the tendency to cause sedation, cough suppression, respiratory depression, and hypoxemia.42
Epidural analgesia (EDA) is a method whereby narcotics, anesthetic agents, or combinations thereof are introduced into the spinal epidural space at the thoracic or lumbar level to provide regional analgesia. This is accomplished by introduction of a polyvinyl catheter into the epidural space and delivery of agents by either a bolus, continuous infusion or, more recently, a demand system.32,39,45–50
The major advantage of EDA is its apparent effectiveness in the absence of sedation.32,39,45-50 EDA has been shown to result in an increased functional residual capacity, lung compliance, and vital capacity; a decreased airway resistance; and increased PO 2.45 Tidal volume is increased and chest wall paradox in flail segments in reduced.28 Patients with EDA generally remain awake and can cooperate with pulmonary toilet.28,47
There are numerous real and theoretical disadvantages to EDA. Insertion may be technically demanding. Epidural anesthetics can cause hypotension, particularly in the face of hypovolemia, and occasional epidural infection.46,47 Epidural hematoma, accidental entry into the spinal canal, and spinal cord trauma can also occur.45 Inadvertent “high block” may lead to respiratory insufficiency. By combining an epidural narcotic with the anesthetic agent, the dose of anesthetic can be decreased and these effects mitigated. However, the narcotic can cause nausea, vomiting, urinary retention, pruritus, and occasionally respiratory depression.28,42,51 The contraindications to EDA may prove problematic in the trauma patient. These include fever, coagulation abnormalities of even minor degrees, and altered mental status. There is some anecdotal concern that the bilateral analgesia effect may mask the symptoms of intra-abdominal injury.52 Finally, nursing intensity in monitoring for the effects of sympathetic block is somewhat more demanding than that for intravenous analgesia.53
Intercostal Nerve Block
Intercostal analgesia or “intercostal nerve block” traditionally involves individual injections of local anesthetic into the posterior component of the intercostal space.45,54–56 Because of segmental overlap of intercostal nerves, it is necessary to induce block above and below any given fractured rib. Blocks of adequate scope have been shown to relieve pain with multiple rib fractures and improve peak expiratory flow rate and volume.57 However, the effect lasts only approximately 6 hours.
As a unilateral block, hypotension is rare, and bladder and lower extremity sensation are preserved. The disadvantages of intercostal block include the need to palpate the fractured ribs for injection, and the need for multiple and repeated injections.45 Local anesthetic toxicity may theoretically occur because of the higher doses needed, and the incidence of pneumothorax increases with the number of ribs blocked.58 Also, inducing block for upper rib fractures is technically difficult because of the proximity of the scapula. Intercostal catheterization and continuous infusion has been successfully used and mitigates the need for multiple injections.43,54 However, the anatomic endpoint of catheter placement, piercing of the “posterior intercostal membrane,” is often unclear, raising the possibility of misplacement.59–61 The full anatomic limits of the spread of intercostal drugs is unclear.60,61
Intrapleural analgesia involves placement of a local anesthetic agent into the pleural space by means of an indwelling catheter.62 The produces a unilateral intercostal nerve block across multiple dermatomes by gravity-dependent retrograde diffusion of agent across the parietal pleura.45 As a unilateral modality, it has advantages similar to intercostal block regarding hypotension and bladder and lower extremity sensation. Successful use of this modality has been reported in blunt thoracic trauma patients.38,63–65
In terms of disadvantages, a significant amount of anesthetic may be lost if a tube thoracostomy is in place, which is often the case with trauma patients.66,67 This can be mitigated by temporary “clamping” of the thoracostomy, which in turn evokes concerns of tension pneumothorax. Conversely, in the absence of a tube thoracostomy, intrapleural catheter placement may cause a pneumothorax. The presence of hemothorax, also common in thoracic trauma patients, may theoretically impair diffusion of anesthetic.68 Because distribution of agent is gravity-dependent, effectiveness also varies with patient position, catheter position, and location of fractured ribs. Diffusion is most widespread in the supine position, which is not optimal for pulmonary function in the trauma patient.45 Conversely, the semiupright position may allow disproportionate diffusion inferiorly and adversely affect diaphragmatic function.69
Thoracic Paravertebral Block
Thoracic paravertebral block involves the administration of a local anesthetic agent in close proximity to the thoracic vertebrae. This can be achieved by intermittent injection, bolus by means of a catheter, or continuous infusion, and produces a unilateral somatic and sympathetic block that extended over multiple dermatomes.31,43,66,70–76
Despite the fact that little recent investigation has been performed with this modality, its theoretical advantages are numerous. It does not require painful palpation of ribs, is not in conflict with the scapula, and is felt by some to be technically easier than epidural anesthesia.74,77 Because there is no risk of spinal cord injury as with EDA, this modality can be instituted on sedated or anesthetized patients. It has few contraindications and requires no special nursing management.73,74 The most common complications are vascular puncture, pleural puncture, and pneumothorax.45 The unilateral nature of the block makes hypotension rare, preserves bladder sensation, and allows monitoring of the lower extremity neurologic examination when necessary. The anatomic location of delivery for the various modalities of regional thoracic analgesia is illustrated in Figure 1.
C. Support for Risk Assessment in Blunt Thoracic Trauma
In 1993, Sariego78 showed that although Trauma Score and Injury Severity Score (ISS) predicted mortality in blunt thoracic trauma, neither identified those survivors who would develop pulmonary complications. Clearly, factors leading to pulmonary sepsis and/or mechanical ventilation set the stage for severe morbidity or mortality. Studies addressing risk assessment in blunt thoracic trauma are tabulated in Table 1 (found online at www.east.org).
Extent of injury to bony thorax
In a very large (n = 692) retrospective Class II series, Svennevig79 identified the presence of four or more rib fractures as an independent predictor of dramatically increased mortality. Patients with three or fewer fractures suffered only a 2.5% mortality, whereas those with four or more had a 19% mortality (p < 0.05). Similarly, in a large (n = 105,000) state registry review (Class III), Lee80 noted a 4% mortality rate for 2,477 patients with three or more rib fractures and a 1% rate for a similar group with two or fewer fractures (p < 0.001). The “two or fewer” fracture group had a statistically similar mortality to the control group in which the patients had no rib fractures.
Finally, Ziegler,81 also in a large retrospective review (n = 711), analyzed mortality in relation to incrementally increasing number of rib fractures. He found a 5% mortality rate with one to two fractures, a 13% mortality rate with three to four fractures, and a 29% mortality rate with seven or more fractures. Analysis of these results did identify an inflection point for increased mortality at four fractures as noted in Figure 2. It should be noted that only 6% of patients had isolated rib fractures, and correction was not made for ISS, which tracked the number of fractures. Consequently, the contribution of the primary chest wall injury to mortality cannot not be isolated reliably.
The salient Class II study was performed by Bergeron and associates82 in 2002. This group prospectively divided 405 patients with rib fractures into a “65 or above age group” and a “less than 65 age group.” The elderly patients had a significantly higher comorbidity rate (61% vs. 8%, p < 0.0001). Their analysis corrected for varying ISS, comorbidity, and a slight difference in mean fracture number. They identified a five-times greater risk of dying in the over 65 age group (9% vs. 19% morality, p < 0.01). This finding is most compelling because the elderly group had a significantly lower ISS despite their higher mortality (p < 0.031).
Finally, an elegant attempt to relate the cumulative or synergistic effects of age and extent of chest wall injury was made by Bulger and colleagues83 in their retrospective (Class II) study of 458 blunt thoracic trauma patients. These authors also divided their population into a customary “65 or older group” and a “younger than 65 group” that were well matched in terms of injury severity. Pneumonia and mortality occurred twice as frequently in the older group (31% vs. 17%, and 22% vs. 10% respectively; both p < 0.01). Similarly, pneumonia and mortality tracked the number of rib fractures in both groups with a mortality odds ratio of 1.2 for each additional fractured rib at any age (p < 0.001). Not surprisingly, the rate of pneumonia increased more rapidly with increasing rib fractures for the elderly group as noted in Figure 3.
The critical finding in this study is that ventilator days, ICU days, hospital length of stay, and mortality increased more rapidly with increasing number of rib fractures for the elderly population. However, this difference was only statistically significant in the midrange of rib fractures, three through six, giving rise to a characteristic curve for these parameters (p ≤ 0.05). This distinctive pattern is illustrated in Figure 4 by the “number of fractures versus mortality” plot.
The authors postulate that this characteristic curve results from the poor tolerance by the elderly for “moderate” levels of injury, which are well tolerated by a younger cohort. At the upper extremes of chest wall injury, both groups do poorly and the curves again approach. All in all, the cumulative effect of age and severity of chest wall injury was powerful. In this study, an elderly person with six rib fractures had a mortality risk of 24% and a pneumonia rate of 35% versus 10% and 20%, respectively, for a younger patient (p < 0.05).
Barnea and colleagues84 retrospectively reviewed 77 elderly (aged ≥ 65 years) with isolated rib fractures. They identified a strong relationship between nonsurvival and the presence of diabetes or congestive heart failure (p = 0.0095 and 0.001). Similarly, Alexander85 retrospectively reviewed 62 elderly patients with isolated rib. Complications occurred in 55% (n = 17) of patients with cardiopulmonary disease (“CPD+” for coronary artery disease or chronic obstructive lung disease) but in only 13% (n = 4) of those without (“CPD–”) (p < 0.05). Mortality occurred only in the CPD+ group (10% n = 3 p < 0.05). Upgrade in level of care was more common in the CPD + group. Length of ICU stay and hospital stay was double in the CPD+ group (p < 0.03).
Conversely, Ziegler81 in a retrospective review of 711 patients, was unable to find a correlation between mortality and the comorbidities of chronic obstructive lung disease (n = 37), diabetes (n = 55), or hypertension (n = 155). There was also no increase in mortality noted for patients with coronary artery disease (n = 116) as defined by a previous myocardial infarction or treatment for angina or for patients with a previous stroke (n = 27). Specific statistical information is not provided in this study.
Concurrent Extrathoracic Injury
The cumulative effect of distant injury on the mortality of blunt thoracic trauma has rarely been specifically addressed. In Svennevig’s79 retrospective, Class II review of 652 blunt trauma patients previously discussed, the presence of one extrathoracic injury did not significantly increase mortality. However, the presence of two extrathoracic injuries increased mortality dramatically, and the highest death rate occurred in the thoracoabdominal injury subgroup (Table 2)
This would not seem surprising, as the ISS has traditionally been accepted as an overall predictor of mortality. However, a number of studies suggest that the ISS may not be a valid predictor of risk of death in the elderly.86–88 Consequently, the incremental effect of distant injury on the mortality of blunt thoracic trauma becomes difficult to assess.
D. Support for Choice of Pain Management Modality 1. Effectiveness of Analgesic Modalities
Thoracic Epidural Analgesia
Studies relating to epidural analgesia are summarized in Table 3 (found online at www.east.org). The greatest recent experience with invasive, regional pain management in the Western world, and in North America in particular, rests with EDA. Nevertheless, there is minimal compelling evidence that EDA improves outcome in trauma patients. Review yielded only one credible study to this end, that by Ullman et al.,39 a landmark Class I review that in 1989 randomized 28 isolated blunt chest trauma patients to receive continuous epidural narcotic or intermittent intravenous injection. The epidural group had significantly less ventilator days (3.1 ± 1.4 vs. 18.3 ± 8.1, p < 0.05), shorter ICU length of stay (5.9 ± 1.5 vs. 18.7 ± 5.3, p < 0.02), and shorter hospital length of stay (14.9 ± 2.2 vs. 47.7 ± 14.6, p < 0.02). The EDA group also had a tracheostomy rate of 7% versus 38% for the control group. Although the sample size was small, the study was adequately powered to the detect the differences indicated.
In an early, Class II study, Gibbons30 in 1973 noted that 27 blunt chest trauma patients treated with lumbar EDA anesthetic required ventilatory support half as frequently as 30 patients who received intravenous narcotic or single-dose intercostal blocks. However, randomization criteria are not specified and there was hesitancy to use thoracic EDA for upper rib fractures at that time. Similarly, in a retrospective study, Wisner40 applied multiple logistic regression analysis to registry data of 465 elderly patients with blunt chest trauma. His group identified the use of EDA as an independent predictor of decreased mortality and pulmonary complications in elderly blunt trauma patients.
Similarly, although EDA is virtually routine in elective thoracic surgery, literature supporting improved outcomes are surprisingly scare for this popular application as well. The solitary Class I study in this field was available only as recently as 2003. In this work, Della Rocca89 showed a 9-day versus 11-day hospital stay for 280 thoracotomy patients who received EDA compared with a similar control group. However, the application of outcome measures from an elective thoracic surgery population to the multiple trauma patient is without validation and conceptually problematic.
Conversely, although quality proof of improved outcome is limited, the evidence that epidural modalities improve subjective pain scores and a variety of pulmonary functions in blunt thoracic trauma patients is abundant and compelling. Four additional Class I studies,37,38,42,44 five Class II studies,29,30,46,50,90 and five Class III studies28,32,49,91,92 document significant improvements in commonly accepted analogue pain scales and such pulmonary parameters as vital capacity, tidal volume, negative inspiratory force, maximum inspiratory flow rate, and minute ventilation. Among salient Class I studies, Moon44 performed a randomized comparison of narcotic/anesthetic epidural with PCA in two well-matched groups (n = 24) of blunt chest trauma patients. The EDA group had a continual increase in maximal inspiratory force (24% from baseline) over the first 3 days, whereas the PCA group had an 18% decrease in the same period. Similarly, initial tidal volume (Vt) for the two groups was not significantly different. However, Vt for the PCA group fell 56% by day 3, whereas that for the EDA group rose by 48%. At the end of this study period, mean Vt was 590 mL for the EDA group versus 327 mL for the PCA group (p < 0.05). Subjective pain scores were similarly dramatically improved (3.8 for EDA vs. 6.2 for PCA, p < 0.05).
Similarly, Mackersie et al.42 randomized 32 multiple rib fracture patients to receive fentanyl by either continuous epidural route or continuous intravenous infusion. Mean vital capacity was dramatically improved in the EDA group versus the intravenous group (5.1 mL/kg vs. 2.8 mL/kg, p < 0.002), as was maximum inspiratory pressure (17 cm H2O vs. 5.3 cm H2O; p < 0.05). In this study, there was no significant change produced in tidal volume, respiratory rate, or minute volume assessed to either method. Although there was a trend toward better improvement in subjective pain scores with EDA, this did not reach statistical significance for the small study group. Similarly, in an early but sizable Class II observational study, Worthley47 treated 147 nonventilated patients with bolus EDA using local anesthetic. A doubling of vital capacity was noted after each dose of the epidural. Nine percent of patients required mechanical ventilation.
The literature derived from elective thoracic surgery is similarly supportive of the benefits of epidural modalities. Four credible Class I studies totaling over 600 patients document very significant improvements in subjective pain control and pulmonary function.89,93–95 One well-designed, Class I study failed to identify any subjective pain score benefits to lumbar epidural fentanyl versus continuous fentanyl infusion.96 However, the study population was small (n = 30), combination epidural anesthetic was not used, and dosing was subjectively titrated for equivalent pain control. Conversely, in a very large prospective review of 2,670 EDA patients and 1,026 intravenous anesthesia (IVA) controls, Flisberg97 noticed dramatic improvement in subjective pain scale.
Other Analgesic Modalities
Little evidence exists for the efficacy of other modalities of invasive, regional analgesia. Ideally, these methods should first be compared with control cases receiving intravenous medication to establish baseline effectiveness. They should then be compared with epidural modalities with which the most experience exists to identify the most effective technique. Studies relating to other modalities of analgesia are summarized in Tables 4 through 6 (found online at www.east.org).
Paravertebral block, as described in greater detail above, is a method in which a bolus injection of anesthetic or a continuous infusion is delivered to the thoracic paravertebral space at the level of rib fractures, producing a unilateral, multilevel, somatic and sympathetic block.72,74 This method is essentially a modality of extrapleural analgesia, as the drug is delivered posterior to the parietal pleura but anterior to the costotransverse ligament near the spine. Although there are a number of anecdotal reports,70,71,98–100 the evidence supporting this modality in trauma patients or general thoracic patients is scant. In a small prospective study (Class II), Gilbert70 administered a single paravertebral anesthetic dose to a mixed group of patients suffering blunt or penetrating thoracic trauma. Vital capacity increased by 65% and respiratory rate decreased by 35%, both to highly significant degrees. Pain scale improved significantly, whereas measures of flow rates (maximum mid-expiratory flow rate and forced expiratory volume in 1 second/forced vital capacity) were unchanged. In a similar Class II, prospective study, Karmakar76 administered continuous paravertebral anesthetic to 15 patients with isolated unilateral rib fractures. There was highly significant (all p < 0.01–0.0001) sustained improvements in analogue pain scores, vital capacity, and peak expiratory flow rate. Interestingly, oxygen saturation (Sao2) and O2 index (PaO2/FIO2 ratio) also improved significantly (p < 0.05).
Extrapleural analgesia is a technique closely related to the paravertebral modality, whereby a catheter is positioned in an extrapleural location and a continuous infusion of local anesthetic is delivered. In a prospective Class I study, Haenal43 administered continuous extrapleural anesthetic to 15 patients with three or more unilateral rib fractures without other injuries. Visual analogue pain scale halved and incentive spirometry doubled. This was significant despite the small study size (p < 0.05). The authors of this study further noted that an anesthesiologist was not used to initiate this therapy at their institution. Similar results have been reported in two Class I101,102 and two Class II103,104 studies in the thoracic surgery literature.
Intrapleural catheters are placed percutaneously in patients with or without chest tubes and used to infuse local anesthetics. They have also been placed through the tracts of in situ tube thoracostomies. Among the salient Class I studies, Gabram104 randomized 42 blunt chest trauma patients to receive systemic narcotics (IVA) or intrapleural anesthetics (IPA). Half the IVA group required crossover to another modality or received mechanical ventilation, whereas this occurred in only 10% of the IPA group (p < 0.05). Changes in pulmonary functions did not reach statistical significance. In a randomized, blinded study, Kottenbelt64 administered intrapleural anesthetic or intrapleural saline to 120 blunt and penetrating trauma patients. Sixty-two percent of the test group but only 15% of the placebo group received satisfactory analgesia as assessed by a visual analogue scale (p < 0.00001). In addition, responders in the treated group had maintenance of their pain relief for a significantly longer period (3.9 hours vs. 0.9 hours, p < 0.005). Pulmonary functions were not assessed. Conversely, in a Class I blinded study of IPA anesthetic versus IPA placebo in 16 blunt trauma patients, Short68 identified no difference in pulmonary function tests, arterial blood gases, subjective pain score, or breakthrough narcotic use. It is noted that the study size was limited. In a similar study, Schneider105 found no benefit to IPA in terms of pain scale, length of stay, or sparing of intravenous narcotics.
Intercostal block was initially performed both by multiple single injections but more recently through a percutaneously placed catheter.31 Murphy54 retrospectively reviewed 57 trauma and general surgery patients treated with multiple intercostal catheter injections of bupivacaine. In this anecdotal, Class III study, patients allowed chest wall palpation and appeared to tolerate physiotherapy better after catheter injection. Analgesia duration was 8 to 12 hours with one dose. All further reviews of intercostal block were embedded in comparative studies and will be considered as such below.
Few comparative studies of the treatment of thoracic pain are to be found in the trauma or general thoracic literature. Shinohara38 performed a small, randomized crossover study examining IPA and EDA in 17 patients with multiple unilateral rib fractures. Subjective pain scores were similar, but because IPA induced a unilateral sympathetic block, blood pressure did not fall with IPA, whereas it did with EDA. However, this difference was not significant however. Luchette and associates37 similarly performed a prospective, randomized comparison of continuous EDA anesthetic versus intermittent IPA anesthetic in 19 blunt thoracic trauma patients. Their epidural patients had significantly less pain at rest and with motion, and this difference continued to widen and was dramatic by day 3. Breakthrough intravenous narcotic use was proportionately different also. Most importantly in this study, tidal volume and negative inspiratory pressure differences were highly significant by day 3 in favor of the epidural route (all p < 0.05). Vital capacity and minute volume were unaffected. These authors concluded that continuous epidural anesthesia was superior to the intrapleural route in terms of pain control and pulmonary function improvement.
Although performed in thoracic surgery patients (n = 40), Bachmann-Mennenga56 carried out an elegant, randomized, four-limb study comparing intercostal block, intrapleural analgesia, thoracic epidural block, and intravenous narcotic. In their study, intercostal and epidural anesthesia produced the greatest pain relief to a high degree of significance (p < 0.01) and had commensurate low levels of narcotic use. Intrapleural block had no narcotic-sparing effect over baseline intravenous analgesia even though catheter placement was confirmed at surgery. It was postulated that the thoracostomy tubes were draining off the anesthetic agent. Although most effective, the epidural route gave the least systemic anesthetic levels. The authors concluded that epidural and intercostal anesthetics constituted the most effective modalities for control of thoracic pain.
Other studies in thoracic surgery patients show preference for intrapleural over intercostal analgesia and paravertebral over intrapleural routes.37,38,55,56,66 However, comparative studies are few and their total patient numbers small. Comparative studies are summarized in Table 7 (found online at www.east.org).
2. Complications and Safety of Analgesic Modalities
A number of sizeable studies have addressed the safety of epidural analgesia in various populations. Scherer106 performed a prospective observational (Class II) study of 1,071 patients in which he reviewed the complication rates but did not address the incidence of expected minor side effects. Patients received epidural narcotic or combination narcotic/analgesic. His group’s findings are listed in Table 8.
Overall treatment-related complications were seen in 37 patients (3.5%). The peripheral nerve damage seen in 0.8% of patients was limited to tingling in various extremities, all of which resolved spontaneously. It is not clear whether some of these may have been related to patient positioning during surgery. There were no sensory or motor deficits, meningitis, or permanent neurologic sequelae. Although 116 patients (10.8%) showed at least one abnormal clotting parameter, there were no clinical hemorrhagic events related to the procedure. One patient experienced respiratory depression temporally related to injection that required intubation. He recovered without sequelae. The authors concluded that EDA was a safe modality with minimal risk of technique-related or pharmacologic complications.
Similarly, Ready107 and colleagues retrospectively reviewed 1,100 postoperative epidural catheters managed outside of an intensive care setting. Narcotics only were used and therefore anesthetic complications such as hypotension were not assessable. These authors noted significant rates of pruritus (25%) and nausea (29%), although neither of these complications was disabling and both were generally managed successfully. The only significant catheter-related problem was dislodgment, which occurred at a rate of 3%. The salient complications noted in this study are summarized in Table 9.
It should be noted for completeness that, as of April 1998, the Food and Drug Administration had recorded 50 spontaneous anecdotal safety reports describing the development of epidural hematomas with the concurrent use of low-molecular-weight heparins (i.e., enoxaparin sodium) and epidural analgesia. The use of these medications for deep venous thrombosis prophylaxis may be a relative contraindication to epidural modalities.108,109
Several studies have attempted to address comparative EDA complication rates against a control of intravenous narcotic.39,42,89,92,93,95–97 These studies are summarized in Table 10 (found online at www.east.org). In general, the smaller studies are often conflicting and fail to identify the same differences in types or rates of complications.39,42,92–94 When considering several larger Class I and II comparative reviews, it is evident that each modality has a unique complication profile but that, in both cases, the rates of significant morbidity are low. Intravenous analgesia tends to have significantly more respiratory depression, central sedative effects, and gastrointestinal effects. Conversely, epidural modalities tend to have more peripheral neurologic effects, pruritus and, when anesthetic agents are used, mild hypotensive effects. Luchette et al. reported significant hypotension with test boluses of lidocaine.37 However, considered together, both modalities have similar, excellent safety profiles.
The single large Class II review of paravertebral analgesia, achieved with local anesthetic agents, prospectively identified a 10% failure rate in 367 cases75 (Table 11, found online at www.east.org). Hypotension requiring treatment occurred in 4.6%. Vascular puncture without morbidity occurred in 3.8% Pleural puncture without pneumothorax occurred in 1.1%, and an additional 0.5% of patients (n = 18) developed a pneumothorax. Some degree of contralateral anesthesia occurred in 1%. There were no instances of entrance into the spinal canal. It should be noted that these cases were accrued from three institutions and therefore represent only modest experience at each center. The time course of the study is not specified and the yearly experience at each institution may be small, thereby accounting for the increased complication and failure rate. Although the authors felt the complication rate was similar to that for epidural anesthesia, other studies have identified lower epidural complication rates at approximately 3%.39,42,89,95 Regardless, no serious complications attributable to PVA were noted in this study. Other studies on paravertebral analgesia are listed in the table. A solitary case of transient Horner syndrome was reported.76 The single small Class II study of the closely related extrapleural analgesia noted no drug or catheter-related complications.43
The majority of small Class I studies addressing the safety of intrapleural catheters identify no significant drug- or catheter-related complications for a total of 151 patients.38,55,64,68,104 However, one prospective observational study of 18 patients noted 11 incorrectly positioned catheters.110 Seven were in lung tissue and three in the chest wall. One tension pneumothorax resulted. The authors postulated that these poor results were experience-dependent. In a small randomized comparative study, Richardson66 noted significant bupivacaine toxicity with intrapleural catheters that did not occur with the paravertebral route. Studies addressing the safety of intrapleural analgesia are summarized in Table 12 (found online at www.east.org). The solitary retrospective, Class III study addressing complication of intermittent intercostal block administered by means of an indwelling catheter identified no catheter- or drug-related complications in 57 patients54 (Table 13, found online at www.east.org).
E. Technical Recommendations Regarding Conduct of Epidural Analgesia
Studies regarding technical recommendations for the conduct of epidural analgesia are summarized in Table 14 (found online at www.east.org).
In 1990, Cicala and colleagues111 compared the effectiveness of a thoracic epidural local anesthetic to a lumbar epidural narcotic in blunt trauma patients. This group found that both modalities were equally effective in decreasing pain scores and that the anesthetic agent was modestly superior in improving pulmonary function tests. The sample size was small (n = 14), although the study was randomized and blinded. The authors theorized that the anesthetic agent benefits pulmonary function by blocking inhibitory neural impulses destined for the diaphragm, thereby improving diaphragmatic function.
The bulk of the information regarding the pharmacology of epidural analgesia arises from the elective thoracic surgical literature. In a randomized blinded study of 53 thoracic surgery patients, Logas showed that epidural narcotic was significantly more effective than anesthetic in subjective pain relief.94 The combination was even more effective. Similarly, other randomized, blinded studies have showed lower pain scores and greater intravenous narcotic sparing with combination therapy as compared with epidural narcotic or anesthetic alone.112,113 Also, it is possible to use lower doses of both agents when used in combination.112
Mode of Infusion
The only study comparing continuous to bolus epidural in trauma patients (blunt and penetrating) was conducted by Kurek and colleagues114 in 1997. In this retrospective study, the continuous infusion method had a slightly though significantly higher complication rate (p < 0.05) than the bolus route. The most common complications with the continuous method were motor blockade (18%), nausea/vomiting (18%), and catheter leaks (12%). For bolus infusions, nausea/vomiting (25%), mental status changes (21%), and local erythema (13%) were most common. There were no serious or permanent complications in either group.
Nursing Environment and Safety Issues
Ready and colleagues53 conducted a large retrospective review (n = 1,106) of a mixed patient population with epidural catheters managed at a general surgical floor level of care. Catheter-related complications occurred in less than 5% of patients. None were serious complications and there were no deaths. A wide variety of nursing protocols exist for monitoring patients receiving epidural anesthetics and narcotics.115,116 All emphasize monitoring of respiratory depression and sedation with epidural narcotic use. They also require frequent assessment of lower extremity motor strength, not only to monitor therapy-related motor blockade caused by anesthetics but for early detection of epidural hematoma and abscess.115,116
In terms of the maximal safe period for use of epidural analgesia, little clear-cut guidance is available in the literature. Consequently, practice is driven by Class III evidence, including a preponderance of custom and expert opinion.28,46,47,91,107,114,117 Clearly, infection is the main time-dependent concern.118 This complication ranges from minimal erythema around the catheter to the rare but devastating complication of epidural abscess with ensuing neurologic injury and possible mortality.118 There have been no studies clearly correlating duration of therapy with incidence of infection. In an early work, Rankin reviewed 50 epidural catheters remaining in situ from 3 to 7.5 days (mean, 5.3 days).46 Four patients (8%) developed induration at the catheter site. The catheters were removed and no further complications occurred. All four catheter tips grew Staphylococcus in culture. One patient may have had an epidural space infection, as evidenced by meningeal symptoms and cerebrospinal fluid leukocytosis. No organisms were recovered. In contrast, Ready retrospectively studied a large, mixed group of 1,106 non-ICU patients receiving epidural analgesia on a general surgery ward.107 The majority of catheters (n = 941 [85%]) remained in situ for 1 to 6 days. However, 147 (13%) were in place for 1 to 2 weeks and 18 (2%) for more than 2 weeks. The longest duration of catheter use was 35 days. No catheters were tunneled. Infection was sought when redness was detected at the catheter site. No infection was identified in this study in 4,343 catheter days. Studies addressing the duration of epidural catheter use are summarized in Table 15 (found online at www.east.org).
In identifying the patients at high risk for morbidity and mortality from blunt chest trauma, outcome clearly worsens with increasing numbers of rib fractures and increasing age. However, identifying a true “inflection point” in the morality curve at which to apply our resources is difficult for either of these parameters. In addition, as a marker of overall injury severity, it is unclear to what extent ameliorating the effects of fractures themselves will improve outcome. Consequently, studies such as those by Svennevig10 that identify rib fractures as an independent predictor of mortality are the most valuable. Nevertheless, it should be remembered that the mortality identified in all studies is real, and attempts to minimize the thoracic contribution to that mortality is appropriate for those patients at significant risk of dying.
Although it is clear that certain analgesic modalities improve subjective pain sensation, the importance of this to recovery, other than in the humanistic sense, is unclear. Although improvement in objective pulmonary function can clearly be documented, the correlation of this to outcome remains somewhat elusive. Just how much improvement in vital capacity is needed to significantly impact ventilator days, or ICU length of stay, or survival? Although most would conceptually agree that improved pulmonary parameters are a good sign in blunt chest injury, the factors affecting outcome, particularly in trauma patients with multiple injuries, are complex and interwoven. Significant populations of isolated chest-injured patients are difficult to mobilize for study. Studies derived from elective thoracic surgery are certainly more available and clearly deal with isolated chest wall pain. However, their validity as models of trauma patients are questionable at best, at least in terms of outcome measures.
As far as effectiveness and complication rates for various modalities, it is reasonable to assume that regional anesthetic techniques, like surgical procedures, have a significant learning curve. Lack of experience with a given modality may contribute to lower success rates and increased complications, thereby negatively impacting on the tendency for future investigation.
Modalities such as intrapleural, extrapleural, or paravertebral analgesia may have greater potential for safety than has been realized and fewer contraindications, which may thus augment their applicability to a trauma population. If efficacy were adequately documented, each of the described modalities offers the promise of its own unique advantages, which would further enhance the armamentarium and pain control flexibility of the trauma surgeon and trauma anesthesiologist. However, the only analgesic modality for which widespread experience exists today in trauma patients is that of epidural administration of narcotics and anesthetics. It is clear that epidural administration of narcotic/analgesic combinations is highly effective in controlling subjective pain and improving pulmonary function. In experienced hands, the rate of complication is minimal and significant morbidity virtually negligible. Contraindications particularly prevalent in the trauma patient, such as slightly abnormal coagulation, spinal fractures, and fever, may limit its use, although the extent to which this occurs is not known. Consequent to the above issues, this group’s recommendations reflect what is known and reasonable regarding identification of those patients at risk from blunt thoracic injury and those analgesic modalities most likely to provide a net positive effect on their outcome.
VI. AREAS FOR FUTURE DEVELOPMENT
On the basis of assessment of current and recent work, the following areas are appropriate for further research:
- Outcome studies regarding epidural analgesia in trauma patients. The effect on primary outcomes of this widely used modality needs to be better defined.
- Outcome studies involving pulmonary function parameters. A correlation needs to be established between improvements in pulmonary function and outcome measures to define specific physiologic goals for therapies.
- Effectiveness/safety of other modalities. Additional investigations need to further evaluate the basic and comparative efficacy of intrapleural, paravertebral/extrapleural, and intercostal modalities. Each of these modalities holds the promise of specific advantages and could extend the flexibility of analgesia if efficacy and safety could be better defined.
- New frontiers. Emerging modalities such as liposomal-encapsulated anesthetic agents offer the potential for safer and more prolonged regional anesthesia.119-121 Trauma surgeons should partner with anesthesiologists to evaluate the applicability of new analgesic modalities for thoracic trauma patients.