Abstract
Objective:
To determine risk factors for osteomyelitis in United States military personnel with combat-related, extremity long bone (humerus, radius, ulna) open fractures.
Design:
Retrospective observational case-control study.
Setting:
U.S. military regional hospital in Germany and tertiary care military hospitals in United States (2003-2009).
Patients/Participants:
Sixty-four patients with open upper extremity fractures who met diagnostic osteomyelitis criteria (medical record review verification) were classified as cases. Ninety-six patients with open upper extremity fractures who did not meet osteomyelitis diagnostic criteria were included as controls.
Intervention:
not applicable.
Main Outcome Measurements:
Multivariable odds ratios (OR; 95% confidence interval [CI]).
Results:
Among patients with surgical implants, osteomyelitis cases had longer time to definitive orthopaedic surgery compared to controls (median: 26 versus 11 days; p<0.001); however, there was no significant difference with timing of radiographic union. Being injured between 2003 and 2006, use of antibiotic beads, Gustilo Anderson [GA] fracture classification (highest with GA-IIIb: [OR: 22.20; CI: 3.60-136.95]), and Orthopaedic Trauma Association Open Fracture Classification skin variable (highest with extensive degloving [OR: 15.61; CI: 3.25-74.86]) were independently associated with osteomyelitis risk. Initial stabilization occurring outside of the combat zone was associated with reduced risk of osteomyelitis.
Conclusions:
Open upper extremity fractures with severe soft-tissue damage have the highest risk of developing osteomyelitis. The associations with injuries sustained 2003-2006 and location of initial stabilization are likely from evolving trauma system recommendations and practice patterns during the timeframe.
INTRODUCTION
Although orthopaedic injuries, such as soft-tissue wounds and closed/open fractures, among combat casualties primarily involve lower extremities, injury to the upper extremities are also often reported.1–5 A retrospective analysis of wounded French military personnel at a combat support hospital in Afghanistan reported that 20% of patients sustained combat-related upper extremity trauma, with soft-tissue injuries being predominant (50%), followed by open fractures (36%) and traumatic amputations (8.5%).6 Between 2005 and 2009, the incidence of upper extremity fractures among combat casualties with musculoskeletal trauma was 1.24 per 1000 deployed personnel per year.3
As with lower extremities, open fractures of upper extremity long bones may be complicated by deep infections or osteomyelitis.7–11 Among 664 combat casualties admitted to orthopaedic service at a U.S. military hospital between 2003 and 2006, 103 (15%) were diagnosed with osteomyelitis and 32 (5%) involved upper extremities.7 High infection rates have also been reported in civilian studies. A large-scale analysis of 3430 patients with humeral shaft fractures (40% open fractures) found that 3% of those with open reduction and internal fixation (ORIF) and 1.5% with intramedullary nailing (IMN) developed an infection.12 Deep infections were also diagnosed in 1.5% of 285 open upper extremity fractures (radius/ulna and humerus) in a prospective cohort analysis.8
When characteristics of 415 open fractures were examined, higher fracture severity, increased injury severity score, and a delay of >8 hours for debridement/irrigation were associated with a risk of infection. When the dataset was separated by extremity site (129 upper and 286 lower extremities; 12% with infections in both groups), lower extremity open fractures retained the same predictive factors; however, no variables were significantly associated with risk of upper extremity infection.10 Muscle damage, wound contamination, and local circulation have also been associated with risk of infection in an analysis of 394 open fractures (including 89 upper extremity long bones).13
Regarding open fractures, prior studies have demonstrated that the combination of multiple factors likely impact the prognosis of patients, including occurrence of infectious complications. Nevertheless, these studies have predominantly focused on the lower extremities with less information available for open upper extremity long bone fractures.8,14–22 Our objective was to retrospectively examine data over a six-year period to investigate risk factors for the development of osteomyelitis in combat-wounded individuals with open fractures of upper extremity long bones.
METHODS
Study Population
Information from U.S. military personnel who sustained deployment-related traumatic orthopaedic injuries between March 19, 2003 and December 31, 2009 and transitioned through Landstuhl Regional Medical Center (LRMC; Germany) to a participating U.S. hospital were examined for inclusion in a retrospective case-control analysis. Participating military hospitals in the United States were Walter Reed Army Medical Center, National Naval Medical Center, and Brooke Army Medical Center. The study was approved by the Institutional Review Board of the Uniformed Services University of the Health Sciences (Bethesda, MD).
Case-Control Classification
Information in the Military Health System Data Repository was surveyed for patients with International Classification of Diseases (9th revision, Clinical Modification) codes related to upper extremity trauma to identify patients for the retrospective analysis. Records from Department of Defense Trauma Registry and Military Health System pharmacy, laboratory, and radiology were also reviewed for relevant information. The study population was restricted to patients with open fractures of upper extremity long bones. Medical records of potential subjects were independently reviewed by an infectious disease clinician and orthopaedic surgeon to verify classification as an osteomyelitis case using standardized clinical diagnostic criteria. Patients were included in the control population if they sustained an open fracture of the upper extremity long bones and did not meet diagnostic criteria for an osteomyelitis.
Open fractures were retrospectively categorized by the orthopaedic surgeon using a modified version of the Gustilo-Anderson (GA) classification system23 involving three established grades: GA-I, GA-II, GA-III, as well as an additional category for traumatic and early surgical transhumeral/transradial amputations (THA/TRA). The Orthopaedic Trauma Association (OTA) Open Fracture Classification (OFC) scheme was also used to categorize open fractures with regards to skin, muscle, arterial, bone loss, and contamination variables.24,25
All patients classified as an osteomyelitis case were evaluated using published Centers for Disease Control and Prevention, National Healthcare Safety Network (NHSN) diagnosis grading of definite, probable, or possible.26 A definite osteomyelitis diagnosis required either a positive bone culture or evidence of bone infection on direct examination during a surgical procedure or histopathological examination. Probable osteomyelitis diagnoses were defined as occurrence of ≥2 clinical signs/symptoms (>38°C temperature; localized swelling, heat, and/or tenderness at site; and drainage at site) in addition to organisms cultured from blood specimens or radiographic evidence of infection. A possible osteomyelitis diagnosis required environmental contamination at time of injury, organism growth from deep wound tissue, and evidence of local or systemic inflammation. In our analysis, patients with a fracture nonunion during follow-up examination and evidence of systemic inflammation were also classified as a possible osteomyelitis.
Statistical Analysis
Characteristics between osteomyelitis cases and controls were compared using chi-square testing (or Fisher’s exact test) for categorical variables and Wilcoxon rank-sum test for continuous variables. Potential risk factors for osteomyelitis were assessed in logistic regression models. Covariates with a p-value ≤0.2 in the univariable model were examined for inclusion in the multivariable model. The final multivariable risk factor model was determined based on stepwise, backward, and forward model selection. Variables with a p-value <0.05 were retained in the multivariable model.
RESULTS
Study Population
A total of 64 osteomyelitis cases and 96 controls were identified and included in the analysis (Table 1). The study population consisted primarily of young (median 24 years of age) men (96%) who served in support of combat operations in Iraq (92%), with the most common injury pattern being via a blast mechanism (64%). Approximately 38% of the study population had moderate injuries, while 36% had severe or life-threatening injuries. Tobacco use was recorded for 43% of patients (46% and 41% of cases and controls, respectively). One patient in the control population died.
Table 1.
Demographics and Injury Characteristics, No. (%), of Wounded Military Personnel with Open Fractures of the Upper Extremity Long Bones by Osteomyelitis Status
| Total (N=160) | Osteomyelitis Cases (N=64) | Controls (N=96) | P-value | |
|---|---|---|---|---|
| Male | 153 (95.6) | 60 (93.8) | 93 (96.9) | 0.439 |
| Age at time of injury, median (IQR) | 24 (21-28) | 23 (21-27) | 25 (21-30) | 0.413 |
| History of tobacco use1 | 58 (43.3) | 26 (46.4) | 32 (41.0) | 0.597 |
| Operational theater | 0.077 | |||
| Afghanistan | 13 (8.1) | 2 (3.1) | 11 (11.5) | |
| Iraq | 147 (91.9) | 62 (96.9) | 85 (88.5) | |
| Time period | 0.101 | |||
| 2003-2006 | 94 (58.8) | 43 (67.2) | 51 (53.1) | |
| 2007-2009 | 66 (41.3) | 21 (32.8) | 45 (46.9) | |
| Branch of service | 0.350 | |||
| Army | 122 (76.3) | 52 (81.3) | 70 (72.9) | |
| Marine | 30 (18.8) | 9 (14.1) | 21 (21.9) | |
| Air Force | 5 (3.1) | 1 (1.6) | 4 (4.2) | |
| Navy | 3 (1.9) | 2 (3.1) | 1 (1.0) | |
| Blast mechanism of injury | 103 (64.4) | 43 (67.2) | 60 (62.5) | 0.615 |
| Blood product transfusions within first 24 hours | 0.116 | |||
| None or missing units | 105 (65.6) | 38 (59.4) | 67 (69.8) | |
| 1-9 units | 32 (20.0) | 18 (28.1) | 14 (14.6) | |
| ≥10 units | 23 (14.4) | 8 (12.5) | 15 (15.6) | |
| Injury severity score | 0.247 | |||
| 0-9 (mild) | 41 (25.6) | 21 (32.8) | 20 (20.8) | |
| 10-15 (moderate) | 61 (38.1) | 22 (34.4) | 39 (40.6) | |
| ≥16 (severe to life-threatening) | 58 (36.3) | 21 (32.8) | 37 (38.5) | |
| Death | 1 (0.6) | 0 | 1 (1.0) | NA |
IQR - interquartile range
Tobacco history was not known for 26 patients (8 cases and 18 controls). Percentages are based on total minus unknown.
Clinical and Open Fracture Characteristics
Among the osteomyelitis cases, 6 patients met the criteria for a definite or probable diagnosis with the remaining 58 patients being classified as a possible osteomyelitis. The definite/probable cases had an increased proportion of purulent drainage or necrotic soft-tissue at the infection site compared to possible cases (67% versus 22%; p=0.038); however, there were no other significant differences between the groups with regards to clinical indicators. In particular, the maximum white blood cell count was comparable between the groups with a median of 13.5 (interquartile range [IQR]: 10.1-19.5) for definite/probable patients and 16.8 (IQR: 11.8-25.1) for possible patients (p=0.417). Among definite/probable patients, the median C-reactive protein level was 9.0 mg/L (IQR: 6.7-9.4) compared to a median of 10.9 mg/L (IQR: 5.1-19.0; p=0.165) for possible patients. Lastly, the median erythrocyte sedimentation rate was 97 (IQR: 90-98) for the definite/probable group and 106 (IQR: 75-120; p=0.165) for the possible group.
Fractures of the humerus were more common among the osteomyelitis cases (58% versus 31% in controls; p=0.002; Table 2). Moreover, a significantly higher proportion of osteomyelitis cases had a fracture classification of GA-IIIb or higher (38% versus 5%) and THA/TRAs (8% versus 4%) compared to controls (p<0.001; Table 2). With regards to the OTA OFC skin, muscle, arterial, bone loss, and contamination variables, cases had more extensive degloving, muscle loss and/or dead muscle, injury with distal ischemia, segmental bone loss, contamination imbedded in bone and deep soft-tissue, and high-risk environmental conditions (p<0.05).
Table 2.
Characteristics, No. (%), of Open Fractures of the Upper Extremity Long Bones Sustained by Military Personnel
| Total (N=160) | Osteomyelitis Cases (N=64) | Controls (N=96) | P-value | |
|---|---|---|---|---|
| Open Fracture Site1 | ||||
| Humerus | 65 (41.9) | 36 (58.1) | 29 (31.2) | 0.002 |
| Radius/Ulna | 95 (61.3) | 30 (48.4) | 65 (69.9) | 0.011 |
| Fracture Class2 | <0.001 | |||
| GA-I | 2 (1.3) | 0 | 2 (2.1) | |
| GA-II | 32 (20.0) | 5 (7.8) | 27 (28.1) | |
| GA-IIIa | 77 (48.1) | 29 (45.3) | 48 (50.0) | |
| GA-IIIb | 16 (10.0) | 12 (18.8) | 4 (4.2) | |
| GA-IIIc/THA/TRA3 | 22 (13.8) | 17 (26.6) | 5 (5.2) | |
| Open fracture not otherwise specified | 11 (6.9) | 1 (1.6) | 10 (10.4) | |
| OTA OFC: Skin4 | <0.001 | |||
| Can be approximated | 80 (52.6) | 17 (27.0) | 63 (70.8) | |
| Cannot be approximated | 42 (27.6) | 25 (39.7) | 17 (19.1) | |
| Extensive degloving | 30 (19.7) | 21 (33.3) | 9 (10.1) | |
| OTA OFC: Muscle,5 | <0.001 | |||
| Grade I | 45 (35.4) | 7 (13.0) | 38 (52.1) | |
| Grade II | 60 (47.2) | 32 (59.3) | 28 (38.4) | |
| Grade III | 22 (17.3) | 15 (27.8) | 7 (9.6) | |
| OTA OFC: Arterial6 | <0.001 | |||
| No injury | 129 (81.1) | 43 (67.2) | 86 (90.5) | |
| Injury without ischemia | 8 (5.0) | 5 (7.8) | 3 (3.2) | |
| Injury with distal ischemia | 22 (13.8) | 16 (25.0) | 6 (6.3) | |
| OTA OFC: Bone loss7 | <0.001 | |||
| None | 63 (40.1) | 15 (23.4) | 48 (51.6) | |
| Bone missing/devascularized with contact by proximal/distal fragments | 59 (37.6) | 28 (43.8) | 31 (33.3) | |
| Segmental bone loss | 35 (22.3) | 21 (32.8) | 14 (15.1) | |
| OTA OFA: Contamination8 | 0.031 | |||
| None or minimal | 9 (5.8) | 1 (1.6) | 8 (8.8) | |
| Surface contamination | 26 (16.8) | 6 (9.4) | 20 (22.0) | |
| Embedded in bone / deep soft-tissue | 64 (41.3) | 31 (48.4) | 33 (36.3) | |
| High-risk environmental conditions | 56 (36.1) | 26 (40.6) | 30 (33.0) |
OTA OFC - Orthopaedic Trauma Association Open Fracture Classification; THA – Transhumeral amputation; TRA – Transradial amputation
Five patients have both upper and lower upper extremity fractures (4 cases and 1 control), so number (and percentages) will sum to more than total. Five patients are missing data (2 cases and 3 controls). Percentages based on total minus missing.
Based on a modified Gustilo-Anderson (GA) classification.23
Amputation prior to infection.
There are 8 patients with unknown classifications (1 case and 7 controls). Percentages based on total minus unknown.
Grade I - no muscle in area, no appreciable muscle necrosis, some muscle injury with intact function; Grade II - loss of muscle but function remains, some localized necrosis requiring excision, intact muscle-tendon unit; Grade III - dead muscle, loss of function, partial/complete excision, complete disruption of unit, defect does not approximate. There are 33 patients with unknown classifications (10 cases and 23 controls). Percentages based on total minus unknown.
There is 1 control patient with an unknown classification. Percentages based on total minus unknown.
There are 3 control patients with unknown classifications. Percentages based on total minus unknown.
There are 5 control patients with unknown classifications. Percentages based on total minus unknown.
Management and Outcomes
Significantly more osteomyelitis cases had their initial stabilization occur prior to medical evacuation from the combat theater (69% versus 39% in controls; p=0.004 Table 3). There was no statistical difference related to the occurrence of internal fixation; however, significantly more osteomyelitis cases had external fixation (73% versus 41%; p<0.001). Moreover, a significantly higher proportion of cases had external fixation as the first type of orthopaedic implant compared to controls (66% versus 37%; p<0.001). Antibiotic beads, predominantly vancomycin, were more frequently administered with osteomyelitis cases (39% versus 11%; p<0.001). Among patients with surgical implants, osteomyelitis cases had a significantly longer time following injury to definitive orthopaedic surgery (median of 26 days; interquartile range [IQR]: 13-222 days compared to 11 days; IQR: 7-24 days; p<0.001). Time to radiographic union was not significantly different between the groups.
Table 3.
Management and Outcomes of Military Personnel with and without Osteomyelitis of the Upper Extremity Long Bones, No. (%)
| Total (N=160) | Osteomyelitis Cases (N=64) | Controls (N=96) | P-value | |
|---|---|---|---|---|
| First orthopaedic implant | <0.001 | |||
| External fixation | 77 (48.1) | 42 (65.6) | 35 (36.5) | |
| Internal fixation | 47 (29.4) | 9 (14.1) | 38 (39.6) | |
| None | 36 (22.5) | 13 (20.3) | 23 (24.0) | |
| Any external fixation of upper extremity | 86 (53.8) | 47 (73.4) | 39 (40.6) | <0.001 |
| Any internal fixation of upper extremity | 119 (74.4) | 52 (81.3) | 67 (69.8) | 0.139 |
| Level of care for initial stabilization1 | 0.004 | |||
| Within combat zone | 65 (51.2) | 36 (69.2) | 29 (38.7) | |
| LRMC (Germany) | 13 (10.2) | 4 (7.7) | 9 (12.0) | |
| U.S. hospital | 49 (38.6) | 12 (23.1) | 37 (49.3) | |
| Presence of foreign body at fracture site2 | 0.182 | |||
| Fragment with orthopaedic implant | 59 (37.1) | 26 (41.3) | 33 (34.4) | |
| Fragment only | 22 (13.8) | 5 (7.9) | 17 (17.7) | |
| Orthopaedic implant only | 68 (42.8) | 26 (41.3) | 42 (43.8) | |
| No foreign body | 10 (6.3) | 6 (9.5) | 4 (4.2) | |
| Time to definitive orthopaedic surgery | 0.005 | |||
| Final stabilization <30 days | 86 (53.8) | 27 (42.2) | 59 (61.5) | |
| Final stabilization 30 to <60 days | 4 (2.5) | 2 (3.1) | 2 (2.1) | |
| Final stabilization 2 - 6 months | 14 (8.8) | 6 (9.4) | 8 (8.3) | |
| Final stabilization >6 months | 23 (14.4) | 17 (26.6) | 6 (6.3) | |
| No orthopaedic implant | 33 (20.6) | 12 (18.8) | 21 (21.9) | |
| Median days to radiographic union (IQR)3 | 83 (57-132) | 96 (65-136) | 76 (54-120) | 0.106 |
| Time to radiographic union3 | 0.076 | |||
| <6 months | 97 (87.4) | 36 (81.8) | 61 (91.0) | |
| 6 to <9 months | 10 (9.0) | 6 (13.6) | 4 (6.0) | |
| 9 to <12 months | 2 (1.8) | 0 | 2 (3.0) | |
| ≥12 months | 2 (1.8) | 2 (4.5) | 0 | |
| Bone graft2 | 0.046 | |||
| Allograft only | 14 (8.8) | 7 (11.1) | 7 (7.3) | |
| BMP only | 2 (1.3) | 1 (1.6) | 1 (1.0) | |
| Autograft | 11 (6.9) | 7 (11.1) | 4 (4.2) | |
| Combination | 3 (1.9) | 3 (4.8) | 0 | |
| Unspecified graft | 7 (4.4) | 1 (1.6) | 6 (6.3) | |
| Antibiotic beads used4 | 41 (27.2) | 27 (45.8) | 14 (15.2) | <0.001 |
| Vancomycin | 33 (21.9) | 23 (39.0) | 10 (10.9) | |
| Prophylactic antibiotics | 0.639 | |||
| Cefazolin | 30 (18.8) | 12 (18.8) | 18 (18.8) | |
| Other antibiotics | 9 (5.6) | 5 (7.8) | 4 (4.2) | |
| Data not available5 | 121 (75.6) | 47 (73.4) | 74 (77.1) |
BMP – bone morphogenetic proteins; IQR – interquartile range; LRMC – Landstuhl Regional Medical Center
Level of care information is missing for 33 patients (12 cases and 21 controls). Percentages based on total minus missing.
Information is unknown for 1 osteomyelitis case patient. Percentages based on total minus unknown.
Data excludes patients with transhumeral/transradial amputations and missing time to radiographic union (total = 111; osteomyelitis cases = 44; controls = 67). Percentages based on total minus unknown.
Antibiotic bead use information is unknown for 5 cases and 4 controls. Percentages based on total minus unknown.
Antibiotic use within combat support hospitals prior to arrival at Landstuhl Regional Medical Center is unknown
Case-Control Risk Analysis
Fracture classification, receipt of 1-9 units of blood within 24 hours post-injury, use of antibiotic beads, presence of a foreign body fragment at fracture site, and OTA OFC related to skin, muscle, arterial, and bone variables were significantly associated with risk of osteomyelitis in the univariable model (Table 4). Occurrence of initial stabilization outside of the combat zone at either LRMC or a U.S. hospital was protective against osteomyelitis risk.
Table 4.
Logistic Regression Analysis of Risk Factors Associated with Osteomyelitis in Wounded Military Personnel with Open Fractures of the Upper Extremity Long Bones
| Multivariable Model | |||
|---|---|---|---|
| Factors | Univariable OR (95%CI) | GA Class Model1 OR (95% CI) | OTA OFC Model1 OR (95% CI) |
| Tobacco history | 1.44 (0.75-2.79) | – | – |
| Time Period | |||
| 2003-2006 | 1.81 (0.94-3.49) | – | 4.20 (1.44-12.30) |
| 2007-2009 | Reference | – | Reference |
| Blast injury mechanism | 1.23 (0.63-2.39) | – | – |
| Blood product transfusions within first 24 hours | |||
| None or missing units | Reference | – | – |
| 1-9 units | 2.27 (1.01-5.06) | – | – |
| ≥10 units | 0.94 (0.37-2.42) | – | – |
| Injury severity score | |||
| 0-9 (mild) | Reference | – | – |
| 10-15 (moderate) | 0.54 (0.24-1.20) | – | – |
| ≥16 (severe to life-threatening) | 0.54 (0.24-1.22) | – | – |
| Fracture classification2 | |||
| GA-I/II | Reference | Reference | – |
| GA-IIIa | 3.50 (1.22-10.06) | 2.54 (0.85-7.60) | – |
| GA-IIIb | 17.39 (3.97-76.17) | 22.20 (3.60-136.95) | – |
| GA-IIIc/THA/TRA | 19.71 (4.98-78.09) | 15.75 (3.87-64.14) | – |
| Use of antibiotic beads | 4.70 (2.19-10.10) | 3.12 (1.31-7.44) | 5.03 (1.67-15.19) |
| OTA OFC: Skin, | |||
| Can be approximated | Reference | – | Reference |
| Cannot be approximated | 5.45 (2.41-12.33) | – | 4.70 (1.57-14.09) |
| Extensive degloving | 7.78 (3.09-19.61) | – | 15.61 (3.25-74.86) |
| OTA OFC: Muscle,3 | |||
| Grade I | Reference | – | – |
| Grade II | 6.20 (2.39-16.08) | – | – |
| Grade III | 10.18 (3.14-33.03) | – | – |
| OTA OFC: Arterial | |||
| No injury | Reference | – | – |
| Injury without ischemia | 3.33 (0.76-14.60) | – | – |
| Injury with distal ischemia | 4.57 (1.75-11.95) | – | – |
| OTA OFC: Bone loss | |||
| None | Reference | – | – |
| Bone missing/devascularized with contact by proximal/distal fragments | 2.89 (1.33-6.26) | – | – |
| Segmental bone loss | 4.48 (1.86-10.81) | – | – |
| OTA OFA: Contamination | |||
| None or minimal | Reference | – | – |
| Surface contamination | 2.40 (0.25-23.24) | – | – |
| Embedded in bone / deep soft-tissue | 7.52 (0.89-63.61) | – | – |
| High-risk environmental conditions | 6.71 (0.79-57.21) | – | – |
| Level of care of initial stabilization | – | – | |
| Within combat theater | Reference | – | Reference |
| LRMC (German hospital) | 0.36 (0.10-1.28) | – | 0.16 (0.03-0.87) |
| U.S. hospital | 0.26 (0.12-0.59) | – | 0.26 (0.09-0.79) |
| Presence of foreign body at fracture site | |||
| Fragment with orthopaedic implant | 0.53 (0.13-2.06) | – | – |
| Fragment only | 0.20 (0.04-0.98) | – | – |
| Orthopaedic implant only | 0.41 (0.11-1.60) | – | – |
| No foreign body | Reference | – | – |
CI – confidence interval; GA – Gustilo-Anderson; LRMC – Landstuhl Regional Medical Center; OR – odds ratio; OTA OFC – Orthopaedic Trauma Association Open Fracture Classification; THA – Transhumeral amputation; TRA – Transradial amputation
GA fracture classification and the OTA OFC variables were highly correlated. The multivariable analysis was run separately after selecting for either the GA fracture classification (GA Class Model) or OTA OFC (OTA OFC Model) to be assessed in the model. Stepwise, backward, and forward model selections were conducted to choose the final multivariable risk factor model.
Using a modified GA classification of open fractures.23 THA/TRA were prior to infection. Patients with open fractures not otherwise specified (1 case and 10 controls) were excluded.
Grade I - no muscle in area, no appreciable muscle necrosis, some muscle injury with intact function; Grade II - loss of muscle but function remains, some localized necrosis requiring excision, intact muscle-tendon unit; Grade III - dead muscle, loss of function, partial/complete excision, complete disruption of unit, defect does not approximate.
The GA fracture and OTA OFC schemes were assessed in separate multivariable models because of high correlation between the variables (Table 4). In the GA fracture classification model, use of antibiotic beads and high fracture severity (i.e., GA-IIIb or greater) were independently associated with osteomyelitis risk. In the OTA OFC model, being injured between 2003 and 2006, use of antibiotic beads, and the OTA OFC skin injury variable were independent predictors, while the highest risk was associated with extensive degloving. Initial stabilization outside of the combat zone remained protective of osteomyelitis risk.
DISCUSSION
Patients with open fractures have a high likelihood of developing infectious complications resulting in hospitalization, surgical procedures, and prolonged antibiotic treatment. Presently, the majority of data examining characteristics, risk factors, and outcomes related to open fractures are focused on the lower extremities, particularly the tibia.8,14–22,27 While our study does have limitations, expansion of the knowledge base related to osteomyelitis risk factors with open upper extremity long bone fractures serves to provide more accurate prognosis of patients. We found that among wounded military personnel with open upper extremity fractures, severe fractures with extensive degloving, use of antibiotic beads, and sustaining an injury between 2003 and 2006 were independent predictors of osteomyelitis. In addition, having initial stabilization occur outside of the combat zone (at LRMC or a U.S. hospital) was protective of osteomyelitis risk.
Increasing severity of fractures according to the GA fracture classification scheme is a recognized risk factor for infection and nonunion.28,29 Specifically, fractures classified as GA-III, as well as transtibial and transfemoral amputations, were significantly associated with risk of osteomyelitis in analyses of open fractures of the tibia and femur.27,30 With open upper extremity fractures, we found that fractures classified as GA-IIIb and above were independent predictors of osteomyelitis risk. It is noteworthy that the highest risk was associated with GA-IIIb fractures (odds ratio: 22.20; 95% confidence interval: 3.60-136.95) and not GA-IIIc or traumatic amputations (THA/TRAs).
Although OTA OFC variables related to muscle damage/loss, distal ischemia, bone loss, and skin were significantly associated with osteomyelitis risk in the univariable model, only the variable for skin was retained in the multivariable model. This is in contrast to prior analyses of the lower extremities, which have shown associations with muscle damage and bone loss.16,27,30 In our analysis, extensive degloving carried the highest risk related to the OTA OFC skin variable (odds ratio: 15.61; 95% confidence interval; 3.25-74.86). Although the blast mechanism of injury was not a risk factor for osteomyelitis in this analysis, the association with extensive degloving can likely be attributed to the high proportion of blast injuries sustained by our study population (67% of cases and 63% of controls), which typically results in complex wounds and mangled extremities characterized by muscle and bone loss.31,32
Initial stabilization outside of the combat zone was associated with a reduced risk of osteomyelitis. Within the combat zone, patients receive care related to damage control and stabilization, including resuscitative care and wound decontamination. Patients may also have multiple surgical procedures prior to medical evacuation from the combat zone related to temporary bone stabilization, as well as debridements and amputations.33,34 Following admission to LRMC (Germany) for intermediate surgical care prior to transfer to the United States, additional wound debridements and irrigations occur as needed. While increased fracture severity and initial stabilization are both independent predictors, these factors are likely related as higher GA fracture severity was associated with stabilization in the combat zone (74% of patients with fractures classified as GA-IIIb or greater received stabilization within combat zone). It is important to note that U.S. military practice patterns were evolving during the study period (2003-2009) as clinical practice guidelines were being developed. This is a likely reason for the association of risk with the time period of 2003-2006.35 In particular, high-pressure irrigation was the standard practice until 2006-2007, when negative pressure wound therapy also became more widely utilized. Use of pulsatile lavage has been identified as a risk factor for osteomyelitis in a meta-analysis (risk ratio of 2.7).36 The association with location of initial stabilization may also be the result of these changing practice patterns related to initial wound care (54% of patients with initial stabilization in combat zone were injured 2003-2006). As high-pressure irrigation was used in the early years of the study, wounds may not have been fully decontaminated before having initial stabilization in the combat zone, leading to a higher potential of infectious complication. In a retrospective analysis of 42 open distal radial fractures (7% with infections),37 a significant association between contamination (e.g., fecal matter, tar, dirt, grass, and gravel) and infection was reported; however, the findings were not assessed in a multivariable model. Time to debridement, number of debridements, and type of fixation were not associated with infection.37 Although occurrence of foreign body at the fracture site and the OTA contamination variable were not significant in our risk factor analysis, the association of location of initial stabilization is potentially related to wound contamination.
Practice patterns also evolved as use of crystalloid products in resuscitation was replaced in favor of a transfusion ratio of 1:1 packed red blood cells and plasma.38 This change may have reduced the risk of late infections, such as osteomyelitis, by decreasing third space fluid; however, greater use of blood component therapy, while effective for life-saving resuscitations, might also increase infection rates through immunosuppression. Thus, further analysis is needed to determine the exact impact of these changing practice patterns.
Regarding antibiotic beads, the association of use with risk of osteomyelitis is likely not a direct causal relationship as use of antibiotic prophylaxis has been demonstrated to reduce the risk of infections with open fractures.39 Rather, the association may be an indicator of the clinical appearance of a wound, overall injury severity, wound size, and geometry, along with the attending surgeon’s suspicion of infection. Reopening the site to remove beads may also impact wound healing. Administration of antibiotics beads are at the discretion of the surgeon based on the wound’s appearance and injury severity. Approximately 81% of antibiotic beads contained vancomycin; however, aminoglycosides were also common (use not well documented). In our analysis, trauma was largely the result of a blast mechanism, resulting in grievous musculoskeletal injuries as characterized by the high proportion of missing or segmental bone loss (38% and 22%, respectively), extensive degloving (20%), and loss of muscle and dead muscle/loss of function (47% and 17%, respectively), according to the OTA OFC scheme. As it is recognized that complex extremity wounds typically have a higher risk of infection,2,34 it is likely that beads were routinely placed in more severe injuries considered at risk for infectious complications.
Type of fixation used with upper extremity fractures has been assessed for association with subsequent infection. With humeral shaft fractures (40% open fractures), IMN has been shown to have a lower proportion of infections compared to ORIF.12 Furthermore, in a meta-analysis, data from seven randomized control trials involving distal radius fractures showed that use of ORIF was associated with a reduced risk of infection compared to external fixation (risk ratio: 0.37; 95% confidence interval: 0.19-0.73).40 In our analysis, the proportion of patients who had ORIF was not significantly different between the cases and controls, while a higher proportion of patients diagnosed with osteomyelitis did have external fixation as the first orthopaedic implant (66% versus 37%; p<0.001), as well as at any time during overall treatment (73% versus 41%; p<0.001). As with antibiotic beads, use of external fixation is likely a hallmark of soft-tissue injury and, thus, should not be associated with osteomyelitis risk. Nevertheless, use of orthopaedic implants was not significant in the univariable model nor was it retained in the multivariable model. This corresponds to the findings of another meta-analysis that did not identify ORIF as a risk factor.36
Our analysis does have limitations inherent to retrospective studies. As 91% of our cases are classified as possible based on NHSN criteria, there is potential inclusion of cases with deep soft-tissue infections without osteomyelitis in our analysis. The determination of diagnostic classifications was based on retrospective review of available medical records. As such, operative descriptions sometimes lacked specificity required to meet criteria for a definite osteomyelitis, owing to the complexity of combat trauma-related open upper extremity long bone fractures or amputations. A recent consensus document has also proposed a new definition for fracture-related infections that warrants consideration.41 Due to the low number of patients with open upper extremity long bone fractures, we were unable to match the case and control populations. Detailed data on the timing and administration of post-trauma antibiotics within the combat zone prior to medical evacuation are lacking. Information on patient-level data related to practice patterns and initial surgical care following injury is also unavailable for a portion of our patients. Lastly, 64% of patients in our population sustained blast-related trauma, resulting in significantly greater injury severity compared to civilian trauma. Therefore, our findings may not be applicable to general civilian trauma.
Overall, osteomyelitis risk is highest among wounded military personnel with open fractures of GA-IIIb or higher, resulting in extensive degloving. Although specific information about the type of internal fixation and timing of surgical implants were not included in this analysis, they could contribute to osteomyelitis risk and thus warrant examination. Future research should also include assessment of the risk of osteomyelitis recurrence with regard to wound severity, microbiology, early initial and late antibiotic therapy, and early surgical care.
Acknowledgments:
The Trauma Infectious Disease Outcomes Study group includes Anuradha Ganesan, MD, Amy Weintrob, MD, Joseph R. Hsu, MD, Denise Bennett, MS, William Bradley, MS, Dan Lu, MS, Lauren Greenberg, MPH, and Jiahong Xu, MS, MPH. We are indebted to our study team of clinical coordinators, microbiology technicians, data managers, clinical site managers, and administrative support personnel for their tireless hours to ensure the success of this project. We thank the U.S. Army Institute of Surgical Research/Joint Trauma System and the Navy Marine Corps Public Health Center for their support in providing data. We also wish to thank M. Leigh Carson for her assistance in preparing the manuscript.
Conflicts of Interest and Sources of Funding: The authors have no conflicts of interest to declare. This study (IDCRP-044) was conducted by the Infectious Disease Clinical Research Program, a DoD program executed through the Uniformed Services University of the Health Sciences, Department of Preventive Medicine and Biostatistics through a cooperative agreement with The Henry M. Jackson Foundation for the Advancement of Military Medicine, Inc. (HJF). This project is funded by the National Institute of Allergy and Infectious Diseases, National Institutes of Health, under Inter-Agency Agreement Y1-AI-5072, and Department of the Navy under the Wounded, Ill, and Injured Program.
Footnotes
Publisher's Disclaimer: Disclaimer: The views expressed are those of the authors and do not reflect the official views or policies of the Uniformed Services University of the Health Sciences, the Henry M. Jackson Foundation for the Advancement of Military Medicine, Inc., the National Institutes of Health or the Department of Health and Human Services, Brooke Army Medical Center, Walter Reed National Military Medical Center, U.S. Army Medical Department, the U.S. Army Office of the Surgeon General, the Department of Defense or the Departments of the Army, Navy or Air Force. Mention of trade names, commercial products, or organizations does not imply endorsement by the United States Government.
REFERENCES
- 1.Andersen CR, D’Alleyrand MP, Swiontkowski MF, et al. Extremity War Injuries VIII: Sequelae of Combat Injuries. J Am Acad Orthop Surg. 2014;22:57–62. [DOI] [PubMed] [Google Scholar]
- 2.Brown KV, Murray CK, Clasper JC. Infectious complications of combat-related mangled extremity injuries in the British military. J Trauma. 2010;69:S109–S115. [DOI] [PubMed] [Google Scholar]
- 3.Belmont PJ Jr., McCriskin BJ, Hsiao MS, et al. The nature and incidence of musculoskeletal combat wounds in iraq and afghanistan (2005-2009). J Orthop Trauma. 2013;27:e107–e113. [DOI] [PubMed] [Google Scholar]
- 4.Owens BD, Kragh JF Jr., Wenke JC, et al. Combat wounds in Operation Iraqi Freedom and Operation Enduring Freedom. J Trauma. 2008;64:295–299. [DOI] [PubMed] [Google Scholar]
- 5.Owens BD, Kragh JF Jr., Macaitis J, et al. Characterization of extremity wounds in Operation Iraqi Freedom and Operation Enduring Freedom. J Orthop Trauma. 2007;21:254–257. [DOI] [PubMed] [Google Scholar]
- 6.Mathieu L, Bertani A, Gaillard C, et al. Wartime upper extremity injuries: experience from the Kabul International Airport combat support hospital. Chir Main. 2014;33:183–188. [DOI] [PubMed] [Google Scholar]
- 7.Yun HC, Branstetter JG, Murray CK. Osteomyelitis in military personnel wounded in Iraq and Afghanistan. J Trauma. 2008;64:S163–168. [DOI] [PubMed] [Google Scholar]
- 8.Weber D, Dulai SK, Bergman J, et al. Time to initial operative treatment following open fracture does not impact development of deep infection: a prospective cohort study of 736 subjects. J Orthop Trauma. 2014;28:613–619. [DOI] [PubMed] [Google Scholar]
- 9.Gupta A, Lakhiani C, Lim BH, et al. Free tissue transfer to the traumatized upper extremity: risk factors for postoperative complications in 282 cases. J Plast Reconstr Aesthet Surg. 2015;68:1184–1190. [DOI] [PubMed] [Google Scholar]
- 10.Malhotra AK, Goldberg S, Graham J, et al. Open extremity fractures: impact of delay in operative debridement and irrigation. J Trauma Acute Care Surg. 2014;76:1201–1207. [DOI] [PubMed] [Google Scholar]
- 11.Yi PH, Weening AA, Shin SR, et al. Injury patterns and outcomes of open fractures of the proximal ulna do not differ from closed fractures. Clin Orthop Relat Res. 2014;472:2100–2104. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Gottschalk MB, Carpenter W, Hiza E, et al. Humeral shaft fracture fixation: incidence rates and complications as reported by American Board of Orthopaedic Surgery Part II Candidates. J Bone Joint Surg Am. 2016;98:e71. [DOI] [PubMed] [Google Scholar]
- 13.Yokoyama K, Itoman M, Nakamura K, et al. New scoring system predicting the occurrence of deep infection in open upper and lower extremity fractures: efficacy in retrospective re-scoring. Arch Orthop Trauma Surg. 2009;129:469–474. [DOI] [PubMed] [Google Scholar]
- 14.Napierala MA, Rivera JC, Burns TC, et al. Infection reduces return-to-duty rates for soldiers with Type III open tibia fractures. J Trauma Acute Care Surg. 2014;77:S194–S197. [DOI] [PubMed] [Google Scholar]
- 15.Penn-Barwell JG, Bennett PM, Fries CA, et al. Severe open tibial fractures in combat trauma: management and preliminary outcomes. Bone Joint J. 2013;95-B(1):101–105. [DOI] [PubMed] [Google Scholar]
- 16.Penn-Barwell JG, Bennett PM, Mortiboy DE, et al. Factors influencing infection in 10 years of battlefield open tibia fractures. Strategies Trauma Limb Reconstr. 2016;11:13–18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Lack WD, Karunakar MA, Angerame MR, et al. Type III open tibia fractures: immediate antibiotic prophylaxis minimizes infection. J Orthop Trauma. 2015;29:1–6. [DOI] [PubMed] [Google Scholar]
- 18.Thakore RV, Francois EL, Nwosu SK, et al. The Gustilo-Anderson classification system as predictor of nonunion and infection in open tibia fractures. Eur J Trauma Emerg Surg. 2017;43:651–656. [DOI] [PubMed] [Google Scholar]
- 19.Duyos OA, Beaton-Comulada D, Davila-Parrilla A, et al. Management of open tibial shaft fractures: does the timing of surgery affect outcomes? J Am Acad Orthop Surg. 2017;25:230–238. [DOI] [PubMed] [Google Scholar]
- 20.Bennett PM, Sargeant ID, Myatt RW, et al. The management and outcome of open fractures of the femur sustained on the battlefield over a ten-year period. Bone Joint J. 2015;97-B:842–846. [DOI] [PubMed] [Google Scholar]
- 21.Srour M, Inaba K, Okoye O, et al. Prospective evaluation of treatment of open fractures: effect of time to irrigation and debridement. JAMA Surg. 2015;150:332–336. [DOI] [PubMed] [Google Scholar]
- 22.Rajasekaran S, Naresh Babu J, Dheenadhayalan J, et al. A score for predicting salvage and outcome in Gustilo type-IIIA and type-IIIB open tibial fractures. J Bone Joint Surg Br. 2006;88:1351–1360. [DOI] [PubMed] [Google Scholar]
- 23.Gustilo RB, Anderson JT. Prevention of infection in the treatment of one thousand and twenty-five open fractures of long bones: retrospective and prospective analyses. J Bone Joint Surg Am. 1976;58:453–458. [PubMed] [Google Scholar]
- 24.Marsh JL, Slongo TF, Agel J, et al. Fracture and dislocation classification compendium - 2007: Orthopaedic Trauma Association classification, database and outcomes committee. J Orthop Trauma. 2007;21:S1–S133. [DOI] [PubMed] [Google Scholar]
- 25.Orthopaedic Trauma Association: Open Fracture Study Group. A new classification scheme for open fractures. J Orthop Trauma. 2010;24:457–464. [DOI] [PubMed] [Google Scholar]
- 26.Centers for Disease Control and Prevention. CDC/NHSN Surveillance Definitions for Specific Types of Infections. 2019. http://www.cdc.gov/nhsn/pdfs/pscmanual/17pscnosinfdef_current.pdf Accessed January 4, 2019.
- 27.Tribble DR, Lewandowski LR, Potter BK, et al. Osteomyelitis risk factors related to combat trauma open tibia fractures: a case-control analysis. J Orthop Trauma. 2018;32:e344–e353. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Bowen TR, Widmaier JC. Host classification predicts infection after open fracture. Clin Orthop Relat Res. 2005:205–211. [DOI] [PubMed] [Google Scholar]
- 29.Westgeest J, Weber D, Dulai SK, et al. Factors associated with development of nonunion or delayed healing after an open long bone fracture: a prospective cohort study of 736 subjects. J Orthop Trauma. 2016;30:149–155. [DOI] [PubMed] [Google Scholar]
- 30.Lewandowski LR, Potter BK, Murray CK, et al. Osteomyelitis risk factors related to combat trauma open femur ffrfafctures: A Case-Control Analysis. J Orthop Trauma. 2019;33:e110–e119. [DOI] [PubMed] [Google Scholar]
- 31.Belmont PJ, Schoenfeld AJ, Goodman G. Epidemiology of combat wounds in Operation Iraqi Freedom and Operation Enduring Freedom: orthopaedic burden of disease. J Surg Orthop Adv. 2010;19:2–7. [PubMed] [Google Scholar]
- 32.Ficke JR, Eastridge BJ, Butler F, et al. Dismounted complex blast injury report of the Army Dismounted Complex Blast Injury Task Force. J Trauma Acute Care Surg. 2012;73:S520–S534. [Google Scholar]
- 33.Mathieu L, Bertani A, Gaillard C, et al. Surgical management of combat-related upper extremity injuries. Chir Main. 2014;33:174–182. [DOI] [PubMed] [Google Scholar]
- 34.Murray CK, Hsu JR, Solomkin JS, et al. Prevention and management of infections associated with combat-related extremity injuries. J Trauma. 2008;64:S239–S251. [DOI] [PubMed] [Google Scholar]
- 35.Hospenthal DR, Murray CK, Andersen RC, et al. Guidelines for the prevention of infection after combat-related injuries. J Trauma. 2008;64:S211–S220. [DOI] [PubMed] [Google Scholar]
- 36.Kortram K, Bezstarosti H, Metsemakers WJ, et al. Risk factors for infectious complications after open fractures; a systematic review and meta-analysis. Int Orthop. 2017;41:1965–1982. [DOI] [PubMed] [Google Scholar]
- 37.Glueck DA, Charoglu CP, Lawton JN. Factors associated with infection following open distal radius fractures. Hand (N Y). 2009;4:330–334. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Langan NR, Eckert M, Martin MJ. Changing patterns of in-hospital deaths following implementation of damage control resuscitation practices in US forward military treatment facilities. JAMA Surg. 2014;149:904–912. [DOI] [PubMed] [Google Scholar]
- 39.Morgenstern M, Vallejo A, McNally MA, et al. The effect of local antibiotic prophylaxis when treating open limb fractures. Bone Joint Res. 2018. 7:447–456. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Esposito J, Schemitsch EH, Saccone M, et al. External fixation versus open reduction with plate fixation for distal radius fractures: a meta-analysis of randomised controlled trials. Injury. 2013;44:409–416. [DOI] [PubMed] [Google Scholar]
- 41.Metsemakers WJ, Morgenstern M, McNally MA, et al. Fracture-related infection: a consensus on definition from an international expert group. Injury. 2018;49:505–510. [DOI] [PubMed] [Google Scholar]