Journal Logo

Adult Respiratory Distress Syndrome, Pneumonia, and Mortality following Thoracic Injury and a Femoral Fracture Treated Either with Intramedullary Nailing with Reaming or with a Plate. A Comparative Study*

BOSSE, MICHAEL J., M.D.†, CHARLOTTE, NORTH CAROLINA; MACKENZIE, ELLEN J., PH.D.‡, BALTIMORE; RIEMER, BARRY L., M.D.§, PITTSBURGH, PENNSYLVANIA; BRUMBACK, ROBERT J., M.D.¶; MCCARTHY, MELISSA L., M.S.‡; BURGESS, ANDREW R., M.D.¶; GENS, DAVID R., M.D.¶; YASUI, YUTAKA, PH.D.#, BALTIMORE, MARYLAND

Article
Free

Multiply injured patients (an Injury Severity Score of 17 points or more) who were admitted to one of two level-I regional trauma centers between 1983 and 1994 because of a fracture of the femoral shaft with a thoracic injury (an Abbreviated Injury Scale score of 2 points or more) or without a thoracic injury were studied retrospectively. The patient populations and the protocols for the treatment of trauma were similar at the two centers; however, the centers differed with regard to the technique that was used for acute stabilization of the fracture of the femoral shaft. At Center I intramedullary nailing with reaming was used in 217 (95 per cent) of the 229 patients, whereas at Center II a plate was used in 206 (92 per cent) of the 224 patients. This difference was used to investigate the effect of acute femoral reaming on the occurrence of adult respiratory distress syndrome in multiply injured patients who had a chest injury.Three groups of patients were evaluated: those who had both a fracture of the femur and a thoracic injury, those who had a fracture of the femur but no thoracic injury, and those who had a thoracic injury without a fracture of the femur or the tibia. The third group was studied at each center to determine if there was a difference between the institutions with regard to the rate of adult respiratory distress syndrome. Patients who had diabetes, chronic obstructive pulmonary disease, asthma, hepatic or renal failure, or an immunosuppressive condition were excluded from the study.The records were abstracted to determine the Injury Severity Score, Abbreviated Injury Scale score, and Glasgow Coma Score for each patient. Requirements for fluid resuscitation were calculated for the first twenty-four hours; these included the number of units of packed red blood cells, fresh-frozen plasma, and platelets that were transfused and the volume of crystalloid that was used. The duration of intubation, the duration of hospitalization, and the occurence of adverse outcomes (death, multiple organ failure, adult respiratory distress syndrome, pneumonia, and pulmonary embolism) were determined for each patient. The groups of patients were analyzed as a whole and then were stratified into subgroups (according to whether or not they had a thoracic injury and whether the Injury Severity Score was less than 30 points or 30 points or more) to determine if the type of fixation of the femoral fracture affected the rate of adult respiratory distress syndrome or mortality. Logistic regression models were used to analyze the data.The over-all occurrence of adult respiratory distress syndrome in the 453 patients who had a femoral fracture was only 2 per cent (ten patients). The rates of adult respiratory distress syndrome for the patients who had a thoracic injury but no femoral fracture (eight [6 per cent] of 129 patients at Center I, compared with ten [8 per cent] of 125 patients at Center II) did not differ between centers, suggesting that the institutions were comparable in their treatment of multiply injured patients. The occurrence of adult respiratory distress syndrome in the patients who had a femoral fracture without a thoracic injury did not differ substantially according to whether the fracture had been treated with a nail (118 patients) or a plate (114 patients). Likewise, the frequency of adult respiratory distress syndrome, pneumonia, pulmonary embolism, failure of multiple organs, or death for the patients who had a femoral fracture and a thoracic injury was similar regardless of whether nailing with reaming (117 patients) or a plate (104 patients) had been used.The use of intramedullary nailing with reaming for acute stabilization of fractures of the femur in multiply injured patients who have a thoracic injury without a major comorbid disease does not appear to increase the occurrence of adult respiratory distress syndrome, pulmonary embolism, failure of multiple organs, pneumonia, or death.

†Department of Orthopaedic Surgery, Carolinas Medical Center, P.O. Box 32861, Charlotte, North Carolina 28262.

‡Center for Injury Research and Policy, School of Hygiene and Public Health, Johns Hopkins University, 624 North Broadway, Baltimore, Maryland 21205.

§420 East North Avenue, 401, Pittsburgh, Pennsylvania 15212.

¶R Adams Cowley Shock-Trauma Center, 22 South Greene Street, T3R59, Baltimore, Maryland 21201.

#Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, 1124 Columbia Street, MP-702, Seattle, Washington 98104.

Early stabilization of fractures of the femoral shaft is the preferred treatment at most major trauma centers, and this practice appears to have contributed to decreased rates of morbidity and mortality for multiply injured patients3,6,8,12,20. Johnson et al. reported a significant increase in the occurrence of adult respiratory distress syndrome in association with a delay in the operative stabilization of femoral fractures in patients who had multiple injuries (fifteen [75 per cent] of twenty patients who had delayed treatment, compared with six [17 per cent] of thirty-five who were managed acutely; p < 0.001). Bone et al.5 conducted a prospective, randomized study of patients who had a fracture of the femoral shaft that was treated either acutely or after a delay. They found that the occurrence of pulmonary complications, including adult respiratory distress syndrome, fat embolism, and pneumonia was higher; the duration of hospitalization was longer; and the number of days in the intensive-care unit was greater for patients with multiple injuries who had delayed fixation of the fracture. Behrman et al. reported that delayed fixation was associated with a significant increase in the occurrence of pulmonary shunt (two of eight patients who had early fixation, compared with twenty-one of twenty-five who had late fixation; p < 0.005) in patients who had an Injury Severity Score of more than 36 points; it also was associated with a significant increase in the occurrence of pneumonia (two of twenty-five patients who had early fixation, compared with seven of twenty-three who had late fixation; p < 0.03) in patients who were older than fifty years. The numbers of days in the intensive-care unit and in the hospital also were increased for patients who had delayed fixation.

Intramedullary nailing is the most commonly used method of fixation of fractures of the femoral shaft, and intramedullary reaming is a component of the operative procedure in most patients. In recent clinical and basic-science reports, it was proposed that intramedullary nailing with reaming adversely affects the lungs by embolizing marrow elements14-18,23. Pape et al.17 found that intramedullary nails inserted with reaming caused additional damage to the lungs in sheep that also had severe shock and pulmonary contusions. Winquist and Bogosian found elevated intramedullary pressures (a mean of 360 milligrams of mercury [47.99 kilopascals]) during preparation of the canal of intact femora being treated with closed shortening procedures, and, like Pell et al., they detected fat embolism with use of transesophageal echocardiography during the reaming process. In a retrospective study, Pape et al.15 concluded that patients in whom a femoral fracture had been treated acutely with intramedullary nailing with reaming in the presence of severe thoracic injury had a higher occurrence of adult respiratory distress syndrome than did similar patients managed in a delayed fashion (eight [33 per cent] of twenty-four, compared with two [8 per cent] of twenty-six; p < 0.05). Those authors deduced that, in patients who have a thoracic injury, acute nailing with reaming may potentiate the development of adult respiratory distress syndrome through marrow embolization and subsequent inflammatory reactions. They recommended either delayed intramedullary nailing or immediate fixation with a technique not necessitating intramedullary reaming for multiply injured patients who have a thoracic injury15.

The conclusions and recommendations of Pape et al.15, however, were based on a small series of patients from a single institution. Furthermore, their patients were assigned to immediate or delayed treatment on the basis of uncontrolled criteria determined by the surgeons, and patients who died from either head injury or hemorrhagic shock were excluded from the study. This methodology increased a potential bias in the comparison of the treatment groups.

The purpose of the current retrospective study was to test the hypothesis that early intramedullary nailing with reaming adversely affects pulmonary function in multiply injured patients. The study was designed to determine the rate of adult respiratory distress syndrome associated with femoral fractures and thoracic injury at two large trauma centers and to use the differences in the fixation techniques—intramedullary nailing with reaming or use of a plate—between the centers to determine if the rate of adult respiratory distress syndrome was lower when a femoral fracture associated with a thoracic injury was treated without intramedullary nailing with reaming.

Back to Top | Article Outline

Materials and Methods

The null hypothesis of this study was that the occurrence of adult respiratory distress syndrome, pulmonary embolism, pneumonia, or multiple organ failure following acute intramedullary nailing with reaming for a femoral fracture associated with a thoracic injury does not differ from the occurrence following treatment with a plate. The hypothesis was tested with use of a retrospective cohort design, with pulmonary complications compared between two groups of multiply injured patients who had sustained a thoracic injury and a femoral fracture but were managed at two different level-I trauma centers. The two centers, while similar in their protocols for early fixation of fractures of long bones and delivery of perioperative critical care, differ with regard to their preferred method of fixation. At Center I (the R Adams Cowley Shock-Trauma Center), fractures of the femoral shaft routinely are fixed with use of intramedullary nailing with reaming7. At Center II (Allegheny General Hospital), acute fixation with a femoral plate is preferred19. This difference in the fixation technique between the institutions allows a comparison of the two methods. To ensure that any observed differences in outcomes could not be attributed to differences in treatment protocols other than the method of fracture fixation, outcomes also were compared between the two institutions for multiply injured patients who had a femoral fracture but no thoracic injury as well as for multiply injured patients who had a thoracic injury but no femoral fracture. After controlling for the severity of the injury, pre-existing medical conditions, and age, we hypothesized that there would be no difference in the rates of adult respiratory distress syndrome between the two institutions for these two subgroups of patients.

Back to Top | Article Outline

Study Population

Three groups of patients who were admitted for the treatment of blunt multisystem injury (an Injury Severity Score2 of 17 points or more) were identified at each trauma center. Group A consisted of patients who had both a femoral fracture and a thoracic injury; Group B, those who had a femoral fracture but no thoracic injury; and Group C, those who had a thoracic injury but no femoral or tibial fracture. Thoracic injuries were defined as injuries to the thorax with an Abbreviated Injury Scale score1 of 2 points or more. Fractures of the femur were defined as only those amenable to treatment with a first-generation locking intramedullary nail and to stabilization within twenty-four hours after the injury. This group included all fractures of the shaft and most subtrochanteric and supracondylar fractures. The decision as to whether the fracture could be treated with nailing was made by an orthopaedic surgeon on the basis of the medical record and the radiographs. Patients who had had a previous fracture of the ipsilateral femur were excluded from the study.

In addition to meeting group-specific criteria, all patients had to have been admitted between July 1, 1983, and June 30, 1994, and to have been between the ages of fifteen and fifty-four years at the time of the injury. Patients were excluded from the study if the medical record contained a diagnosis of insulin-dependent diabetes, acquired immunodeficiency syndrome, chronic use of steroids, renal or hepatic failure, or lung disease (chronic obstructive pulmonary disease, emphysema, tuberculosis, or asthma). The criteria for exclusion were applied to each record before the Abbreviated Injury Scale score was determined and before what we considered to be adverse outcomes for the purpose of this study (death, adult respiratory distress syndrome, multiple organ failure, pulmonary embolism, or pneumonia) were identified.

Back to Top | Article Outline

Procedures

Patients who were eligible for the study initially were identified from the trauma registries maintained at the two institutions. Subsequent review of the complete medical record by a research nurse was necessary to confirm eligibility. All patients who met the criteria for Group A or Group B were included in the study. A comparable number of patients were identified for Group C by selecting every eighteenth patient who met that group's criteria within the same year of discharge.

The medical records were abstracted by one research nurse at each institution. Both research nurses were hired and trained specifically for this study. To ensure consistency and accuracy in the coding of severity and the identification of adverse outcomes, the two research nurses participated together in a two-day training course. Training included a review of the operational definitions of adverse outcomes (to be discussed) as well as the conventions used in assigning Abbreviated Injury Scale scores to individual injuries. To ensure additional consistency in the abstracting and in the assignment of the Abbreviated Injury Scale scores, the records for the first twenty-five patients studied at each institution were independently abstracted and scores were assigned by a third research nurse who is an expert in this type of coding and who led the training session for the study. At both institutions, agreement between the study research nurse and the expert research nurse was more than 90 per cent for all items. Differences in coding were reviewed with each research nurse.

The information obtained from the medical record included demographic data on the patient; the Abbreviated Injury Scale score for each injury sustained as well as the over-all Injury Severity Score; the location and type of femoral fracture and the method of fixation; the Glasgow Coma Score21; the respiratory rate and the systolic blood pressure at the time of admission to the hospital; the requirements for fluid resuscitation during the first twenty-four hours and for the entire stay in the hospital; the duration and parameters of intubation and mechanical respiration; the duration in the hospital and in the intensive-care unit; and the onset and duration of adverse outcomes, including adult respiratory distress syndrome, pneumonia, pulmonary embolism, fat embolism, multiple organ failure, and death.

Back to Top | Article Outline

Definitions of Adverse Outcomes

Of primary interest in this study was the comparison of the rates of adult respiratory distress syndrome. Other pulmonary related conditions, however, may be associated with intramedullary nailing with reaming. The rates of multiple organ failure, pneumonia, infection, and pulmonary embolism therefore also were recorded and were compared among study groups. Specific definitions were developed to standardize the coding of adverse outcomes. The duration of use of the respirator was recorded as the number of days that a patient needed to use a ventilator with a positive end respiratory pressure of more than zero. Pneumonia was defined as a temperature of more than 38.0 degrees Celsius; a white blood-cell count of more than 10,000; a radiograph of the chest showing a persistent infiltrate; and a sputum gram stain showing less than ten epithelial cells per high-power field, more than twenty-five white blood cells per high-power field, and organisms. Generalized sepsis was defined as a temperature of more than 38.5 degrees Celsius, a leukocytosis, and a positive bacterial blood culture. Adult respiratory distress syndrome was defined as a PaO2/FiO2 ratio of less than 200 for at least five consecutive days4 and bilateral diffuse infiltrates seen on the radiograph of the chest in the absence of pneumonia and cardiogenic pulmonary edema. Adult respiratory distress syndrome was considered to be related to the fixation of the femoral fracture if symptoms and signs developed within five days after the operation. Multiple organ failure was defined as a Multiple Organ Failure Score of 4 points or more. Each organ system had to demonstrate dysfunction for at least three days concomitantly13.

Back to Top | Article Outline

Analysis of the Data

The characteristics of the patients and the occurrence of adverse outcomes were compared within each of the three study groups. The significance of the differences in proportions was established with use of chi-square analysis. A p value of 0.05 was considered significant. Student t tests were used to compare means. The observed differences in the occurrence of adverse outcomes together with the appropriate 95 per cent confidence intervals are reported. When differences in the nature and severity of the injuries sustained by patients in different treatment groups were identified, multiple-regression analyses were performed to take these differences into account when comparing the occurrences of adverse outcomes.

The nature and severity of the injuries were summarized in two different ways. First, we developed models that included both Injury Severity Scores and Glasgow Coma Scores together with an indicator for the presence of bilateral thoracic injury. Recognizing the limitations of the Injury Severity Score in reflecting the severity of multiple injuries to the same region of the body, we also developed a model in which the components of the Anatomic Profile9 together with the Glasgow Coma Score were used as independent variables in the regression model. The three component scores of the Anatomic Profile equal the square root of the sum of the squares of the Abbreviated Injury Scale scores of greater than 2 points for injuries of the head and the spinal cord, the thorax and the anterior aspect of the neck, and all other regions of the body.

Back to Top | Article Outline

Results

Seven hundred and seven patients were included in the study: 221 in Group A (a femoral fracture with a thoracic injury), 232 in Group B (a femoral fracture without a thoracic injury), and 254 in Group C (a thoracic injury without a femoral or tibial fracture). An additional thirty-five patients who had a femoral fracture met the criteria for inclusion in Group A or Group B but were managed with an external fixator. The external fixator was used at both institutions for wound care and as a minimally invasive technique of stabilization for patients who were in extremis and whose condition was deemed too unstable for either immediate nailing with reaming or immediate fixation with a plate but who needed stabilization of the femoral fracture for critical-care management. Twenty of these thirty-five patients had an open fracture. These patients were excluded from the study.

Most (217) of the 229 patients in Groups A and B who were managed at Center I had intramedullary nailing with reaming, whereas most (206) of the 224 who were managed at Center II were managed with a plate (Table I). The method of fixation of the fracture depended on the surgeon but generally was in accordance with the institutional treatment philosophy. Reasons for crossovers in technique at each institution could not be determined, and they were attributed to vagaries in the preference of the surgeon. The small group of thirty patients who were managed with the procedure performed predominantly at the other center were classified according to the method of fixation (and not according to the institution) in the analysis. The analysis was repeated with these thirty patients excluded, but the conclusions remained the same.

TABLE I - DISTRIBUTION OF THE SEVEN HUNDRED AND SEVEN PATIENTS ACCORDING TO GROUP AND TYPE OF TREATMENT
Group A: Femoral Fracture with Thoracic Injury Group B: Femoral Fracture without Thoracic Injury Group C: Thoracic Injury without Femoral Fracture
Center I Center II Center I Center II Center I Center II
Nail 112 5 105 13
Plate 6 98 6 108
      Total 118 103 111 121 129 125

Back to Top | Article Outline

Group C (Thoracic Injury Only)

The distributions of age and gender in Group C did not differ between the two institutions (Table II). The patients also had comparable lengths of stay in the hospital (eighteen days for those managed at Center I and nineteen days for those managed at Center II), and they were managed with a respirator for a comparable number of days (seven days for those managed at Center I and six days for those managed at Center II). Although the two groups of patients were also similar with regard to the over-all Injury Severity Score (Fig. 1), a significantly higher percentage of those who were managed at Center II had bilateral thoracic injury (thirty-five [28 per cent] of 125, compared with thirteen [10 per cent] of 129; p < 0.0001). The mean Glasgow Coma Score was significantly higher at Center I (p < 0.0001), and a higher percentage of patients at Center I had severe abdominal injuries (an Abbreviated Injury Scale score of 4 points or more; p = 0.04) and higher over-all requirements for fluid resuscitation (p = 0.0001).

Fig

Fig

TABLE II - DATA ON GROUP C (THORACIC INJURY ONLY) ACCORDING TO CENTER
Center I (N = 129) Center II (N = 125) P Value Mean Difference (95 Per Cent Confidence Interval)
Variable
      Male gender (no. of patients) 87 (67%) 87 (70%) >0.5
      Mean age (yrs.) 29 29 >0.5
      Mean Injury Severity Score2 (points) 31 30 >0.5
      Injury Severity Score2≥40 points 21 (16%) 17 (14%) >0.5
            (no. of patients)
      Abbreviated Injury Scale score1 ≥ 4 points
            (no. of patients)
                  Head 29 (22%) 40 (32%) 0.09
                  Thorax 59 (46%) 54 (43%) >0.5
                  Abdomen 23 (18%) 11 (9%) 0.04
      Bilateral involvement of chest (no. of patients) 13 (10%) 35 (28%) <0.0001
      Mean Glasgow Coma Score21 (points) 13 11 <0.0001
      Mean requirement for resuscitation
            Red blood cells (units) 6 3 0.02
            Platelets (units) 2 2 0.47
            Fresh-frozen plasma (units) 5 2 <0.0001
            Crystalloids (ml) 7671 4801 0.0001
      Mean duration on respirator (days) 7 6 >0.5
      Mean hospital stay (days) 18 19 >0.5
Adverse outcome (no. of patients)
    Adult respiratory distress syndrome 8 (6%) 10 (8%) >0.5 -0.02 (-0.08–0.05)
    Multiple organ failure 9 (7%) 1 (1%) 0.02 0.06 (0.02–0.11)
    Multiple organ failure or adult respiratory 12 (9%) 10 (8%) >0.5 0.01 (-0.06–0.08)
        distress syndrome
    Pneumonia 15 (12%) 29 (23%) 0.02 -0.12 (-0.21–0.02)
    Pulmonary embolism 0 (0%) 2 (2%) 0.24 -0.02 (-0.04–0.01)
    Death due to multiple organ failure 5 (4%) 0 (0%) 0.06 0.04 (0.01–0.07)
        or adult respiratory distress syndrome

Comparison of adverse outcomes in Group C revealed some differences between the two institutions with regard to the rates of multiple organ failure and pneumonia; Center I had a higher rate of multiple organ failure, whereas Center II had a higher rate of pneumonia. After controlling for the small differences in the severity of the injuries at the two institutions, however, these differences no longer remained significant (p > 0.06). The rates of adult respiratory distress syndrome at the two centers were comparable (eight [6 per cent] of 129 patients at Center I, compared with ten [8 per cent] of 125 patients at Center II). Although one could argue that the current study was not sufficiently large to detect a difference in the rates of adult respiratory distress syndrome as small as the observed difference of 0.02 (power, 10 per cent), it was large enough to detect differences of 0.08 to 0.10 (power, more than 80 per cent). These results provide strong evidence that the treatment of pulmonary injuries resulted in similar outcomes at the two institutions.

Back to Top | Article Outline

Group B (Femoral Fracture without Thoracic Injury)

As with Group C, the age and gender distributions in Group B were similar between the two institutions (Table III). However, there were significant differences in the mix of the severity of the injuries (Fig. 2). Specifically, patients who were managed with intramedullary nailing (primarily those who were managed at Center I) tended to be more severely injured (p < 0.001), tended to stay in the hospital longer (p = 0.01), and were managed with a respirator for a mean of three more days (p = 0.01) than those who were managed with a plate (primarily those who were managed at Center II).

Fig. 2

Fig. 2

TABLE III - DATA ON GROUP B (FEMORAL FRACTURE WITHOUT THORACIC INJURY) ACCORDING TO METHOD OF TREATMENT
Nail (N = 118) Plate (N = 114) P Value Mean Difference (95 Per Cent Confidence Interval)
Variable
      Male gender (no. of patients) 79 (67%) 89 (78%) 0.06
      Mean age (yrs.) 25 27 0.1
      Mean Injury Severity Score2 (points) 26 23 <0.001
      Injury Severity Score2 ≥ 40 points 4 (3%) 0 (0%) 0.12
            (no. of points)
      Abbreviated Injury Scale score1 ≥ 4 points
            (no. of patients)
                  Head 36 (31%) 19 (17%) 0.01
                  Abdomen 13 (11%) 4 (4%) 0.03
      Mean Glasgow Coma Score21 (points) 12 13 0.2
      Mean requirement for resuscitation
            Red blood cells (units) 6 6 >0.5
            Platelets (units) 2 2 >0.5
            Fresh-frozen plasma (units) 4 3 0.19
            Crystalloids (ml) 9150 8727 0.46
      Mean duration on respirator (days) 6 3 0.01
      Mean hospital stay (days) 24 20 0.01
Adverse outcome (no. of patients)
      Adult respiratory distress syndrome
            Femoral fracture* 3 (3%) 1 (1%) >0.5 0.02 (-0.02–0.05)
            Entire admission 4 (3%) 1 (1%) 0.37 0.03 (-0.01–0.06)
      Multiple organ failure 1 (1%) 0 (0%) >0.5 0.01 (-0.01–0.03)
      Multiple organ failure or adult respiratory 4 (3%) 1 (1%) 0.37 0.03 (-0.01–0.06)
            distress syndrome
      Pneumonia 8 (7%) 3 (3%) 0.14 0.04 (-0.01–0.1)
      Pulmonary embolism 0 (0%) 1 (1%) 0.49 -0.01 (-0.03–0.01)
      Death
            Due to multiple organ failure 2 (2%) 1 (1%) >0.5 0 (-0.02–0.02)
                  or adult respiratory distress syndrome
            Due to any cause 3 (3%) 4 (4%) >0.5 -0.01 (-0.05–0.03)
*
Adult respiratory distress syndrome developed within five days after fixation of the femoral fracture.

The rates of adult respiratory distress syndrome, multiple organ failure, and pneumonia were all somewhat higher among patients who had been managed with nailing than among those who had been managed with a plate. However, none of these differences was significant (p > 0.41). As mentioned earlier, although the study was not large enough to detect differences as small as those observed, it was large enough to detect differences in the rates of adverse outcomes of at least 0.05 (power, more than 70 per cent). The mean length of stay (p = 0.01) and the mean number of days of treatment with a respirator (p = 0.01) were significantly higher for patients who had been managed with nailing with reaming than for those who had been managed with a plate, without adjusting for group differences in the severity of the injuries. After controlling for the difference in the severity of the injuries between the two groups, the effect of the nailing on either the mean length of stay (p > 0.2) or the mean number of days on the respirator (p = 0.06) failed to reach significance. The difference between the groups with regard to the over-all severity of the injuries and, in particular, the presence of a severe abdominal injury (an Abbreviated Injury Scale score of 4 points or more) appears to explain the differences in both the number of days of treatment with a respirator and the total length of stay. In both regression models, these variables were the most significant predictors of outcome (p < 0.01).

Back to Top | Article Outline

Group A (Femoral Fracture with Thoracic Injury)

In Group A, the age and gender distribution were similar between the two institutions (Table IV). However, the patients who were managed with nailing had a higher rate of Injury Severity Scores that were considered to be severe (40 points or more) (twenty-five [21 per cent] of 117) than did those managed with a plate (ten [10 per cent] of 104) (Table IV and Fig. 3). These higher scores appeared to be associated with a higher rate of severe abdominal injuries (fifteen [13 per cent] of the 117 patients who had nailing compared with seven [7 per cent] of the 104 who had fixation with a plate).

Fig. 3

Fig. 3

TABLE IV - DATA ON GROUP A (FEMORAL FRACTURE WITH THORACIC INJURY) ACCORDING TO METHOD OF TREATMENT
Nail (N = 117) Plate (N = 104) P Value Mean Difference (95 Per Cent Confidence Interval)
Variable
      Male gender (no. of patients) 92 (79%) 72 (69%) 0.11
      Mean age (yrs.) 28 28 >0.5
      Mean Injury Severity Score2 (points) 30 28 0.07
      Injury Severity Score2 ≥ 40 points 25 (21%) 10 (10%) 0.02
            (no. of patients)
      Abbreviated Injury Scale score1 ≥ 4 points
            (no. of patients)
                  Head 21 (18%) 19 (18%) >0.5
                  Thorax 26 (22%) 20 (19%) >0.5
                  Abdomen 15 (13%) 7 (7%) 0.13
      Bilateral involvement of chest (no. of patients) 11 (9%) 16 (15%) 0.18
      Mean Glasgow Coma Score21 (points) 13 12 0.34
      Mean requirement for resuscitation
            Red blood cells (units) 8 7 0.42
            Platelets (units) 3 3 >0.5
            Fresh-frozen plasma (units) 5.7 3.3 0.007
            Crystalloids (ml) 10,623 7827 <0.001
      Mean duration on respirator (days) 8 6 0.09
      Mean hospital stay (days) 24 22 0.33
Adverse outcome (no. of patients)
      Adult respiratory distress syndrome
            Femoral fracture* 3 (3%) 1 (1%) >0.5 0.02 (-0.02–0.05)
            Entire admission 3 (3%) 2 (2%) >0.5 0.01 (-0.03–0.05)
      Multiple organ failure 4 (3%) 0 (0%) 0.12 0.03 (0.00–0.07)
      Multiple organ failure or adult respiratory 5 (4%) 2 (2%) 0.45 0.02 (-0.02–0.07)
            distress syndrome
      Pneumonia 12 (10%) 11 (11%) >0.5 0.00 (-0.08–0.08)
      Pulmonary embolism 3 (3%) 3 (3%) >0.5 0.00 (-0.05–0.04)
      Death
            Due to multiple organ failure or adult 1 (1%) 0 (0%) >0.5 0.01 (-0.01–0.03)
                  respiratory distress syndrome
            Due to any cause 3 (3%) 2 (2%) >0.5 0.01 (-0.03–0.05)
*
Adult respiratory distress syndrome developed within five days after fixation of the femoral fracture.

The rates of adverse outcomes for the two treatment groups, however, were similar (Table IV). Specifically, there was no difference in the rates of adult respiratory distress syndrome, pulmonary embolism, or pneumonia. Patients who were managed with nailing had a slightly higher rate of multiple organ failure, but this difference was not significant (p = 0.12). As was true for the comparison of rates in the other two groups, the size of the present study was only large enough to detect differences of 0.05 with sufficient power (more than 80 per cent). However, it is important to note that both subsets of patients in Group A had a low rate of adult respiratory distress syndrome (three [3 per cent] of the 117 who had nailing, compared with one [1 per cent] of the 104 who had fixation with a plate; p > 0.5).

Back to Top | Article Outline

Discussion

Numerous authors have identified a relationship between early fixation of fractures of long bones, particularly femoral fractures, and reductions in the number of days in the intensive-care unit; the number of days in the hospital; and the rates of adult respiratory distress syndrome, pneumonia, infection, and death3,5,6,8,12. Fractures of the femur are most commonly stabilized with intramedullary nailing. Reaming of the medullary canal to accept a nail with an appropriate diameter is a common practice.

Pape et al.15 recently challenged the philosophy of early fixation of femoral fractures with intramedullary nailing with reaming in multiply injured patients. Those authors retrospectively studied patients who had an Injury Severity Score of more than 18 points, a thoracic injury (an Abbreviated Injury Scale score of 2 points or more), and treatment of a femoral fracture with nailing with reaming. There was a higher occurrence of adult respiratory distress syndrome and a higher rate of mortality in association with early fixation than in association with delayed fixation (eight [33 per cent] of twenty-four compared with two [8 per cent] of twenty-five, and five [21 per cent] of twenty-four compared with one [4 per cent] of twenty-five, respectively). Pape et al.15 concluded that, in the presence of a thoracic injury, primary intramedullary nailing with reaming of the femur causes additional pulmonary damage and may trigger the development of adult respiratory distress syndrome. They recommended either a delay in primary internal fixation or, if the femoral fracture were to be treated acutely, the use of temporary external fixation or a different method of internal fixation such as nailing without reaming or plate fixation.

Pape et al.14,16 followed their initial clinical report with a non-randomized clinical study in which they compared the pulmonary function of patients who had a femoral fracture that had been treated with reaming with the function of those managed without reaming. Using a PaO2/FiO2 ratio to assess pulmonary function, they found decreased function in the patients who had been managed with nailing with reaming.

The adverse pulmonary outcomes reported by Pape et al.14-16 had not been recognized at major trauma centers in the United States; however, the observation warranted additional studies to determine if the method of fracture treatment would adversely affect the patient's pulmonary recovery. Charash et al. repeated the study design of Pape et al. with similarly sized groups of patients who had a femoral fracture and a thoracic injury, and they reported contradictory findings. There was a 48 per cent rate of pneumonia in twenty-five patients who were managed with delayed fixation, compared with a 14 per cent rate in fifty-six patients who were managed with early fixation. The over-all rate of pulmonary complications (pneumonia, adult respiratory distress syndrome, fat embolism, and pulmonary embolism) was 56 per cent in the patients who had delayed fixation, compared with 16 per cent in those who had early fixation (p = 0.007). Charash et al. concluded that delayed fixation in patients who have a thoracic injury is associated with a higher rate of pulmonary complications.

Van Os et al. retrospectively studied two groups of multiply injured patients who had a severe thoracic injury with (twenty-seven patients) or without (thirty patients) major fractures of the extremities treated with early fixation. With the numbers available, no significant difference was noted between the two groups with regard to the pulmonary outcomes.

In retrospective series reported on by Johnson et al., Pape et al.15, Charash et al., van Os et al., and Bone et al.6, the results for multiply injured patients managed with early fixation were compared with those for patients who had delayed treatment. The patients were not randomized, and, given that care was provided in centers that advocated early stabilization of long-bone fractures, the patients who were managed with delayed fixation potentially represent a much different subset: those whose condition was too unstable at the time of admission to have safe early operative stabilization.

The basis of the theory of Pape et al.15—that intramedullary nailing with reaming potentiates the development of adult respiratory distress syndrome in patients who have thoracic injury—is the expected effect of embolized bone marrow on pulmonary function. Embolization of marrow products has been well described11,14,18,23; however, the effect of the marrow products on pulmonary function is unclear. Appropriate animal models have been sought to define the impact, if any, of the method of fixation of the femoral fracture on pulmonary function. The basic-science evidence supporting clinically important pulmonary effects of femoral nailing with reaming is inconclusive. While some research has suggested a possible association17,25, other studies have not demonstrated clinically important alterations in pulmonary function10,11,24. The use of a technique without reaming did not reduce the peak intramedullary pressures thought to be related to the embolic phenomenon11.

The reported occurrence of adult respiratory distress syndrome after multiple injury in association with a femoral fracture has varied5,8,12,15,22; Johnson et al. reported it to be as high as 75 per cent (fifteen of twenty) in patients who had delayed fixation of major fractures (femur, pelvis, and spine). The reported differences are related partially to the criteria selected by the authors for the diagnosis of adult respiratory distress syndrome and for attributing the pulmonary dysfunction to the long-bone fracture instead of to an associated pulmonary injury or to other combined systemic factors. In the present study, we used the diagnostic criteria of Pape et al.15 in order to be able to compare our results with their observations. The criteria used to diagnose adult respiratory distress syndrome in the current report are similar to the recommendations of the American-European Consensus Conference on ARDS4, but they are more rigid in one respect: the symptoms in our patients had to be present for at least five consecutive days. The patients who met the criteria and died before the fifth day of symptoms were considered to have had adult respiratory distress syndrome. The pulmonary dysfunction had to have begun before the sixth postoperative day in order to be considered as having a relationship to the fixation of the fracture.

The present clinical study was designed in an attempt to clarify the relationship between the method of fixation of the femoral fracture and the subsequent development of an adverse pulmonary outcome in multiply injured patients who had a thoracic injury. We surmised that, if femoral nailing with reaming provides a bolus of marrow products to the lung and if the marrow products induce an inflammatory cascade that potentiates the development of adult respiratory distress syndrome in patients who have a thoracic injury, then we would see a higher rate of adult respiratory distress syndrome in patients who had been managed with nailing with reaming than in a similar group who had been managed with plate fixation.

Our findings did not confirm those of Pape et al.15. The current study focused only on multiply injured patients who had an acutely treated fracture of the femur. The injury profiles are very similar to those of the patients who were reported on by Pape et al.15 in their initial clinical paper. Assuming that it is the technique for the treatment of the femoral fracture and not the fracture itself, methods of resuscitation, associated injuries, prehospital and postoperative care, or comorbid conditions that determines the eventual development of adult respiratory distress syndrome, we anticipated (but did not find) a difference between the rate of adverse outcomes associated with intramedullary nailing with reaming and that associated with use of a plate in patients who had a thoracic injury. The over-all rate of adult respiratory distress syndrome was only 2 per cent. There were no detectable differences in the rates of adult respiratory distress syndrome, pneumonia, multiple organ failure, pulmonary embolism, or death between the patients who did and those who did not have a thoracic injury regardless of whether they had been managed acutely with nailing or with use of a plate. When the groups were stratified according to the Injury Severity Score, we did not see an increased rate of adult respiratory distress syndrome in the patients who had a thoracic injury and an Injury Severity Score of more than 30 points compared with those who did not have a thoracic injury and a score of more than 30 points.

The current study is limited by its retrospective and non-randomized design. Clearly, a more definitive conclusion would necessitate a prospective, randomized clinical trial with strict criteria for inclusion and exclusion. However, on the basis of the results of the present study, a very large series of patients managed at multiple institutions would be required. The cost of such a trial would be prohibitively high, and the study would be of questionable value on the basis of the preponderance of the evidence accumulated to date.

The strengths of the current study, however, are notable. The conclusions are based on the analysis of a large cohort of patients who had multiple trauma and a femoral fracture with or without an associated thoracic injury. The nurses who abstracted the medical records at the two trauma centers followed the same study protocol to ensure uniform coding of the severity of the injuries and the adverse outcomes. Adverse outcomes were assigned on the basis of current definitions, and the development of adult respiratory distress syndrome was identified with regard to its temporal relationship to the fixation of the femoral fracture. (Adult respiratory distress syndrome that develops at least twenty-one days after femoral fixation and is related to abdominal infection or aspiration cannot be attributed to the method of fixation of the long-bone fracture.) Although it could be argued that the results of the study may be biased because the nurses who reviewed the records for adverse outcomes were not blinded to the severity or treatment of the injuries, we attempted to minimize this bias by hiring and training these individuals specifically for this study. They were not employees of either institution and, to our knowledge, had no vested interest in the results of the study. In addition, as noted earlier, we developed objective criteria for use by the nurses in identifying patients who had an adverse outcome. Any questions regarding the application of the criteria were referred to one of us (D. R. G.).

The study sites are both recognized as high-volume regional level-I trauma centers with expertise in the treatment of fractures of the femur7,19. The centers agreed on general treatment protocols but disagreed on the method of fixation of the fracture; this allowed us to compare the results of nailing with reaming with those of fixation with a plate. This study included only patients who had early fixation, eliminating the potential for selection bias associated with a retrospective comparison of patients who had acute fixation with those who had delayed fixation. Lastly, patients who had a femoral fracture without a thoracic injury were compared with those who had a thoracic injury without a femoral or tibial fracture, to ensure that any difference noted in the patients who had a femoral fracture and a thoracic injury was a result of the method of fixation and not of a difference between the institutions.

Adult respiratory distress syndrome usually is related to a direct or an indirect thoracic injury. The syndrome is recognized as resulting from infection, aspiration, multiple transfusions, intensive resuscitation with fluids, fracture, pulmonary contusion, intensive-care-related pneumonia, closed head injury, toxic inhalation, or a combination of these conditions4,26. Multiple risk factors increase the occurrence of adult respiratory distress syndrome. Clinical studies implicating a specific factor as a cause or catalyst for the development of adult respiratory distress syndrome in patients who have multiple trauma must be able to control for the prehospital care of the patient (airway management, intubation, and tube thoracostomy), the delivery of critical care after admission, resuscitation with fluids, use of blood products, nutritional parameters, preinjury comorbid conditions, age, immune status, and associated major injuries.

This study does not support the hypothesis that intramedullary nailing with reaming in multiply injured patients who have a thoracic injury potentiates the development of adult respiratory distress syndrome. Other major or minor injuries, combinations of these injuries, therapeutic variations, patient-related variables, or bias in the selection of patients might account for the differences that have been reported for other series.

*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. Funds were received in total or partial support of the research or clinical study presented in this article. The primary funding sources were the AO Foundation and AO North America. In addition, support for the data analysis was provided by National Center for Injury Prevention and Control Grant R49/CCR302486.

Investigation performed at the Departments of Orthopaedic Surgery, Carolinas Medical Center, Charlotte, Allegheny General Hospital, Pittsburgh, the R Adams Cowley Shock-Trauma Center, Baltimore, and the Center for Injury Research and Policy, School of Hygiene and Public Health, Johns Hopkins University, Baltimore

Back to Top | Article Outline

References

1. Association for the Advancement of Automotive Medicine: The Abbreviated Injury Scale. 1990 Revision. Des Plaines, Illinois, Association for the Advancement of Automotive Medicine, 1990.
2. Baker, S. P.; O'Neill, B.; Haddon, W., Jr.; and Long, W. B.: The Injury Severity Score: a method for describing patients with multiple injuries and evaluating emergency care. J. Trauma, 14: 187-196, 1974.
3. Behrman, S. W.; Fabian, T. C.; Kudsk, K. A.; and Taylor, J. C.: Improved outcome with femur fractures: early vs. delayed fixation. J. Trauma, 30: 792-797, 1990.
4. Bernard, G. R.; Artigas, A.; Brigham, K. L.; Carlet, J.; Falke, K.; Hudson, L.; Lamy, M.; Legall, J. R.; Morris, A.; and Spragg, R.: The American-European Consensus Conference on ARDS. Definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am. J. Respir. and Crit. Care Med., 149: 818-824, 1994.
5. Bone, L. B.; Johnson, K. D.; Weigelt, J.; and Scheinberg, R.: Early versus delayed stabilization of femoral fractures. A prospective randomized study. J. Bone and Joint Surg., 71-A: 336-340, March 1989.
6. Bone, L. B.; McNamara, K.; Shine, B.; and Border, J.: Mortality in multiple trauma patients with fractures. J. Trauma, 37: 262-264, 1994.
7. Brumback, R. J.; Reilly, J. P.; Poka, A.; Lakatos, R. P.; Bathon, G. H.; and Burgess, A. R.: Intramedullary nailing of femoral shaft fractures. Part I: Decision-making errors with interlocking fixation. J. Bone and Joint Surg., 70-A: 1441-1452, Dec. 1988.
8. Charash, W. E.; Fabian, T. C.; and Croce, M. A.: Delayed surgical fixation of femur fractures is a risk factor for pulmonary failure independent of thoracic trauma. J. Trauma, 37: 667-672, 1994.
9. Copes, W. S.; Champion, H. R.; Sacco, W. J.; Lawnick, M. M.; Gann, D. S.; Gennarelli, T.; MacKenzie, E.; and Schwaitzberg, S.: Progress in characterizing anatomic injury. J. Trauma, 30: 1200-1207, 1990.
10. Duwelius, P. J.; Huckfeldt, R.; Mullins, R. J.; Shiota, T.; Woll, T. S.; Lindsey, K. S.; and Wheeler, D.: The effects of femoral intramedullary reaming on pulmonary function in a sheep lung model. J. Bone and Joint Surg., 79-A: 194-202, Feb 1997.
11. Heim, D.; Regazzoni, P.; Tsakiris, D. A.; Aebi, T.; Schlegel, U.; Marbet, G. A.; and Perren, S. M.: Intramedullary nailing and pulmonary embolism: does unreamed nailing prevent embolization? An in vitro study in rabbits. J. Trauma, 38: 899-906, 1995.
12. Johnson, K. D.; Cadambi, A.; and Seibert, G. B.: Incidence of adult respiratory distress syndrome in patients with multiple musculoskeletal injuries: effect of early operative stabilization of fractures. J. Trauma, 25: 375-384, 1985.
13. Marshall, J. C., and Nathens, A. B.: Multiple organ dysfunction syndrome. In American College of Surgeons Care of the Surgical Patient, edited by D. W. Wilmore, L. Y. Cheung, A. H. Harken, J. W. Holcroft, and J. L. Meakins. Section II, chapter 8, pp. 1-22. New York, Scientific American, 1995.
14. Pape, H. C.; Regel, G.; Dwenger, A.; Sturm, J. A.; and Tscherne, H.: Influence of thoracic trauma and primary femoral intramedullary nailing on the incidence of ARDS in multiple trauma patients. Injury, 24 (Supplement 3): 82-S103, 1993.
15. Pape, H. C.; Auf'm'Kolk, M.; Paffrath, T.; Regel, G.; Sturm, J. A.; and Tscherne, H.: Primary intramedullary femur fixation in multiple trauma patients with associated lung contusion—a cause of posttraumatic ARDS?. J. Trauma, 34: 540-547, 1993.
16. Pape, H. C.; Regel, G.; Dwenger, A.; Krumm, K.; Schweitzer, G.; Krettek, C.; Sturm, J. A.; and Tscherne, H.: Influences of different methods of intramedullary femoral nailing on lung function in patients with multiple trauma. J. Trauma, 35: 709-716, 1993.
17. Pape, H. C.; Dwenger, A.; Regel, G.; Schweitzer, G.; Jonas, M.; Remmers, D.; Krumm, K.; Neumann, C.; Sturm, J. A.; and Tscherne, H.: Pulmonary damage after intramedullary femoral nailing in traumatized sheep: is there an effect from different nailing methods?. J. Trauma, 33: 574-581, 1992.
18. Pell, A. C. H.; Christie, J.; Keating, J. F.; and Sutherland, G. R.: The detection of fat embolism by transoesophageal echocardiography during reamed intramedullary nailing. A study of 24 patients with femoral and tibial fractures. J. Bone and Joint Surg., 75-B(6): 921-925, 1993.
19. Riemer, B. L.; Butterfield, S. L.; Burke, C. J.; and Kelly, J.: Titanium LC/DCP plates and indirect reduction technique fixation of femoral diaphyseal fracture due to high energy blunt trauma [abstract]. J. Orthop. Trauma, 7: 156, 1993.
20. Seibel, R.; LaDuca, J.; Hassett, J. M.; Babikian, G.; Mills, B.; Border, D. O.; and Border, J. R.: Blunt multiple trauma (ISS 36), femur traction, and the pulmonary failure-septic state. Ann. Surg., 202: 283-295, 1985.
21. Teasdale, G., and Jennett, B.: Assessment of coma and impaired consciousness. A practical scale. Lancet, 2: 81-84, 1974.
22. van Os, J. P.; Roumen, R. M.; Schoots, F. J.; Heystraten, F. M.; and Goris, R. J.: Is early osteosynthesis safe in multiple trauma patients with severe thoracic trauma and pulmonary contusion?. J. Trauma, 36: 495-498, 1994.
23. Winquist, R. A., and Bogosian, A. J.: Relationship of intramedullary reamer design to intramedullary pressure and fat embolism. Orthop. Trans., 18: 1022, 1994-1995.
24. Wolinsky, P. R.; Sciadini, M. F.; Parker, R. E.; Plitman, J. D.; Snapper, J. R.; Rutherford, E. J.; Schulman, M.; and Johnson, K. D.: Effects on pulmonary physiology of reamed intramedullary nailing in an open chested sheep model. Orthop. Trans., 19: 151-152, 1995.
25. Wozasek, G. E.; Thurnher, M.; Redl, H.; and Schlag, G.: Pulmonary reaction during intramedullary fracture management in traumatic shock: an experimental study. J. Trauma, 37: 249-254, 1994.
26. Zapol, W. M., and Lemaire, F. [editors]: Adult Respiratory Distress Syndrome. New York, Dekker, 1991.
Copyright © 1997 by The Journal of Bone and Joint Surgery, Incorporated