Abstract
Objective
This study sought to provide a national, descriptive analysis to determine fat embolism syndrome (FES) risk factors, hypothesizing that femur fractures and multiple fractures are associated with an increased risk.
Methods
The Trauma Quality Improvement Program was queried (2010–2016) for patients with extremity fractures. A multivariable logistic regression analysis model was used.
Results
From 324,165 patients, 116 patients (0.04%) were diagnosed with FES. An age ≤30, closed femur fracture, and multiple long bone fractures were associated with an increased risk of FES.
Conclusion
Future research to validate these findings and develop a clinical risk stratification tool appears warranted.
Keywords: Fat embolism syndrome, Extremity fracture, Humerus fracture, Femur fracture, Tibia fracture
1. Introduction
Fat embolism syndrome (FES) is a rare complication of long bone fractures with an incidence ranging from 0.9 to 11%1,4,5,12 in trauma patients and can develop within hours to days post-injury.6,24 FES occurs when adipose tissue passes into the circulation, typically from a fracture, which then migrates and obstructs a vessel, often collecting in the pulmonary or cerebral vasculature.7 While rare, it can lead to acute respiratory distress syndrome (ARDS), respiratory failure, and even brain death secondary to hypoxia and/or massive cerebral fat emboli leading to cerebral edema.2,27 Mechanical ventilation is required in as many as 50% of patients with FES,8 and patients that develop FES have a mortality rate of 5–15%.3,16
The classic triad of respiratory insufficiency, central nervous system dysfunction, and petechial rash is quite rare in patients with FES, but up to 80% of patients have neurologic symptoms and up to 50% have a petechial rash.3,9,10 Diagnosis of FES can be challenging due to the fact that only non-specific clinical and laboratory studies exist,11 such as the presence of fat globules in the blood and urine, or bronchial lavage producing >30% lipid-inclusion-filled alveolar macrophages.3,12,13 Furthermore, these studies are not routinely performed and the laboratory processing may be quite time-consuming. Several different diagnostic criteria for FES have been proposed, including the criteria from Gurd and Wilson. These criteria require one major feature (respiratory insufficiency, cerebral involvement [e.g. altered mentality, confusion, stupor, coma, depression disproportionate to hypoxemia], or petechial rash) and four minor features (pyrexia, tachycardia, jaundice, retinal changes, renal changes, thrombocytopenia, anemia, and elevated erythrocyte sedimentation rate), together with fat macroglobulinemia.14 Therapy for FES is largely supportive consisting of supplemental oxygen and/or mechanical ventilation to offset hypoxia, as well as fluid and/or vasopressor support for circulatory collapse. It is therefore crucial for clinicians to be aware of the greatest risk factors for the development of FES in order to facilitate early diagnosis and prompt supportive care as this may reduce the complications (e.g. hypoxia leading to tissue damage) and mortality associated with FES.15,17
The majority of the published studies regarding FES have been single-center case reports or case series,6,8,11,17,29 which have demonstrated an association of femur fractures and multiple fractures with FES.18,19 In addition, pelvic fractures and closed fractures have also been demonstrated to be associated with a higher rate of FES.21 We sought to provide a nationwide, descriptive analysis on FES in blunt trauma patients, hypothesizing femur fractures and multiple fractures to be associated with an increased risk of FES. In addition, we aimed to evaluate other previously proposed risk factors for FES as well as elucidate additional novel risk factors.
2. Methods
The Trauma Quality Improvement Program (TQIP)22 was queried from 2010 to 2016 for patients ≥18 years-old who suffered blunt trauma resulting in fractures of the humerus, femur, tibia, and patella. These were identified by the appropriate international classification of diseases (ICD) version-9 diagnosis codes. No patients that met the inclusion criteria were excluded from the study. The TQIP database is a large trauma database that collects data from over 800 trauma centers across the United States.22 The primary outcome studied was the rate of FES. The patients were divided into two cohorts: those with and those without FES. Secondary outcomes studied included total hospital length of stay (LOS), intensive care unit (ICU) LOS, ventilator days, and mortality. Additional information collected included the presence of multiple fractures, and complications including acute kidney injury, ARDS, cardiac arrest with cardiopulmonary resuscitation, central line associated bloodstream infection, extremity compartment syndrome, cerebrovascular accident (CVA), unplanned ICU admission, unplanned intubation, pneumonia, deep vein thrombosis, and pulmonary embolism (PE).
We analyzed patients’ demographics and injury profiles, including mechanism of injury and injury severity score (ISS). Additional information including hypotension on admission, concomitant injuries, fracture classification including open or closed, and pre-hospital comorbidities including diabetes, hypertension, congestive heart failure, chronic obstructive pulmonary disease, and cirrhosis was collected. Information regarding fracture treatment was also obtained, however this data is limited by the terminology utilized within the TQIP database (e.g. internal/external fixation, open/closed reduction), which was not designed as an orthopedic surgery database and lacks pertinent details regarding the specific management.
Descriptive statistics were performed for all variables. To compare continuous variables, we used a Mann-Whitney-U test, and these data were reported as means with standard deviation or as medians with interquartile ranges. To study the association between categorical variables, we used a chi-square test, and these data were reported as percentages. A multivariable logistic regression analysis was used to measure the strength of the association between predictor variables and mortality. A hierarchical multivariable logistic regression analysis was used to control for covariates using an odds ratio and associated 95% confidence interval. In the multivariable analysis, the covariates controlled for were multiple fractures, open and closed femur fractures, open and closed humerus fractures, open and closed tibia fractures, pelvic fractures, age ≤30, and non-operative fractures. In addition, we performed a multivariable analysis evaluating serious body region injuries using an Abbreviated Injury Scale (grade >2) for head, spine, abdomen, chest, lower extremity, and upper extremity for risk of development of FES. A p-value <0.05 was considered statistically significant and all p-values were two-sided. The statistical software used was IBM SPSS Statistics for Windows, Version 24 (Armonk, 96 NY: IBM Corp). This study adhered to all pertinent recommendations and standards as outlined in the Strengthening the Reporting of Observational studies in Epidemiology guidelines. Institutional Review Board approval was obtained and informed consent was not possible as the TQIP holds a deidentified dataset. The authors report no proprietary or commercial interest in any product mentioned or concept discussed in this article.
3. Results
3.1. Patient demographics
Out of 324,165 patients in total, 116 (0.04%) suffered from FES. Of FES patients, 101 patients (87.1%) had femur fractures, 42 patients (36.2%) had tibia fractures, and 14 patients (12.1%) had humerus fractures. The FES cohort had a median age of 29 years-old, compared to 59 years-old in the non-FES cohort (p < 0.001). The FES cohort was 70.7% male in comparison to 50.9% male in the non-FES group (p < 0.001). The two cohorts had similar rates of comorbidities including smoking and obesity (p > 0.05). However, the rates of hypertension (13.8% vs. 40.1%, p < 0.001) and dementia (0.9% vs. 5.9%, p = 0.02) were lower in the FES cohort (Table 1). The FES cohort had increased rates of pelvic (20.7% vs. 11.8%, p = 0.003) and rib fractures (26.7% vs. 16.6%, p = 0.004). The FES cohort had higher rates of high speed mechanisms including motor vehicle collisions (39.7% vs. 22.5%, p < 0.001) and motorcycle collisions (22.4% vs. 8.2%, p < 0.001) (Table 2).
Table 1.
Demographics of blunt trauma patients with long-bone extremity fractures.
| Without embolism |
Fat embolism |
||
|---|---|---|---|
| Characteristic | (n = 324,049) | (n = 116) | p-value |
| Age, year, median (IQR) | 59.0 (37, 76) | 29.0 (23, 51) | <0.001 |
| Male, n (%) | 164,802 (50.9%) | 82 (70.7%) | <0.001 |
| ISS, median (IQR) | 9.0 (11, 16) | 12.0 (11, 21) | <0.001 |
| Hypotension on admission, n (%) | 12,711 (4.2%) | 9 (8.5%) | 0.03 |
| Comorbidities, n (%) | |||
| Congestive heart failure | 17,470 (5.4%) | 6 (5.2%) | 0.92 |
| Smoker | 54,047 (16.7%) | 19 (16.4%) | 0.93 |
| Chronic renal failure | 5960 (1.8%) | 1 (0.9%) | 0.43 |
| Diabetes | 48,929 (15.1%) | 14 (12.1%) | 0.36 |
| Hypertension | 129,889 (40.1%) | 16 (13.8%) | <0.001 |
| Chronic obstructive pulmonary disease | 31,501 (9.7%) | 8 (6.9%) | 0.31 |
| Cerebrovascular accident | 10,310 (3.2%) | 5 (4.3%) | 0.49 |
| Cancer, disseminated | 2956 (0.9%) | 0 | 0.30 |
| Dementia | 19,149 (5.9%) | 1 (0.9%) | 0.02 |
| History of myocardial infarction | 5075 (1.6%) | 1 (0.9%) | 0.54 |
| Peripheral vascular disease | 2478 (0.8%) | 0 | 0.34 |
| Psychiatric illness | 24,189 (7.5%) | 6 (5.2%) | 0.35 |
| Obesity | 17,763 (5.5%) | 8 (6.9%) | 0.50 |
| Steroid use | 3174 (1.0%) | 2 (1.7%) | 0.42 |
| Cirrhosis | 2395 (0.7%) | 2 (1.7%) | 0.22 |
| Alcohol use, n (%) | 35,515 (12.1%) | 10 (9.9%) | 0.49 |
| Drug use, n (%) | 38,794 (13.8%) | 23 (24.5%) | 0.003 |
| Race, n (%) | |||
| African-American | 32,939 (11.5%) | 12 (12.8%) | 0.71 |
| Caucasian | 252,773 (78%) | 82 (71%) | 0.06 |
| Asian | 4886 (1.9%) | 2 (2.4%) | 0.75 |
| Hispanic | 23,653 (8.6%) | 12 (12.2%) | 0.20 |
| Teaching university, n (%) | 170,174 (52.5%) | 70 (60.3%) | 0.09 |
| Glasgow coma scale, mean | 14.08 | 13.79 | 0.99 |
| IQR = interquartile range, ISS = injury severity score | |||
Table 2.
Mechanisms and injuries of blunt trauma patients with extremity fractures.
| Without embolism |
Fat embolism |
||
|---|---|---|---|
| Characteristic | (n = 324,049) | (n = 116) | p-value |
| Mechanism, n (%) | |||
| Bicycle accident | 6363 (2.0%) | 0 | 0.13 |
| Motorcycle collision | 26,506 (8.2%) | 26 (22.4%) | <0.001 |
| Pedestrian struck | 22,924 (7.1%) | 7 (6.0%) | 0.66 |
| Fall | 179,512 (55.4%) | 26 (22.4%) | <0.001 |
| Motor vehicle collision | 72,774 (22.5%) | 46 (39.7%) | <0.001 |
| Injury, n (%) | |||
| Traumatic brain injury | 57,532 (17.8%) | 25 (21.6%) | 0.29 |
| Face or neck | 23,696 (7.3%) | 9 (7.8%) | 0.85 |
| Spine | 49, 277 (15.2%) | 24 (20.7%) | 0.10 |
| Rib | 53,909 (16.6%) | 31 (26.7%) | 0.004 |
| Sternum | 8671 (2.7%) | 6 (5.2%) | 0.10 |
| Pelvic fracture | 38,113 (11.8%) | 24 (20.7%) | 0.003 |
| Upper extremity | 82,914 (25.6%) | 39 (33.6%) | 0.047 |
| Lower extremity | 297,345 (91.8%) | 114 (98.3%) | 0.01 |
| Esophagus | 51 (0.02%) | 0 | 0.89 |
| Lung | 46,486 (14.3%) | 35 (30.2%) | <0.001 |
| Heart | 1743 (0.5%) | 4 (3.4%) | <0.001 |
| Stomach | 239 (0.1%) | 0 | 0.77 |
| Small intestine | 2223 (0.7%) | 2 (1.7%) | 0.18 |
| Colorectal | 2385 (0.7%) | 2 (1.7%) | 0.21 |
| Pancreas | 41 (0.01%) | 0 | 0.90 |
| Liver | 11,793 (3.6%) | 9 (7.8%) | 0.02 |
| Spleen | 11,960 (3.7%) | 10 (8.6%) | 0.01 |
| Kidney | 5527 (1.7%) | 4 (3.4%) | 0.15 |
| Other genitourinary | 2614 (0.8%) | 0 | 0.33 |
| Crush | 913 (0.3%) | 1 (0.9%) | 0.24 |
| Burn | 757 (0.2%) | 0 | 0.60 |
The FES cohort had increased rates of closed femur fractures (81.9% vs. 67.6%, p = 0.001). The FES cohort also had increased rates of femur fractures receiving external fixation (12.1% vs. 2.4%, p < 0.001), open reduction and internal fixation (39.7% vs. 28.5%, p = 0.01), and open femur fracture debridement (6.0% vs. 2.8%, p = 0.04). However, the FES cohort had lower rates of non-operative femur fractures (13.9% vs. 32.9%, p < 0.001). The FES cohort had higher rates of open tibia fractures (16.4% vs. 10.6%, p = 0.04) and tibia fractures treated with open reduction and internal fixation (21.6% vs. 14.0%, p = 0.02), compared to the non-FES cohort. The two cohorts had similar rates of closed humerus fractures (9.5% vs. 11.1% in the non-FES group, p > 0.05) and operative humerus fractures (7.8% vs. 7.0% in the non-FES group, p > 0.05). There was a higher rate of multiple fractures in the FES group (44.0% vs. 19.3%, p < 0.001) (Table 3).
Table 3.
Classifications of extremity fractures in blunt trauma patients.
| Without embolism | Fat embolism | ||
|---|---|---|---|
| Characteristic | (n = 324,049) | (n = 116) | p-value |
| Humerus fracture, n (%) | |||
| Closed | 35, 954 (11.1%) | 11 (9.5%) | 0.58 |
| Open | 8153 (2.5%) | 3 (2.6%) | 0.96 |
| External fixation | 935 (0.3%) | 1 (0.9%) | 0.25 |
| Internal fixation without reduction | 279 (0.1%) | 0 | 0.75 |
| Closed reduction without external fixation | 3818 (1.2%) | 2 (1.7%) | 0.59 |
| Closed reduction with external fixation | 871 (0.3%) | 0 | 0.58 |
| Open reduction without fixation | 426 (0.1%) | 0 | 0.70 |
| Open reduction and internal fixation | 18,571 (5.7%) | 8 (6.9%) | 0.59 |
| Debridement | 4882 (1.5%) | 3 (2.6%) | 0.34 |
| All operative | 22,582 (7.0%) | 9 (7.8%) | 0.74 |
| Non-operative | 21,537 (49.5%) | 5 (35.7%) | 0.30 |
| Femur fracture, n (%) | |||
| Closed | 219,136 (67.6%) | 95 (81.9%) | 0.001 |
| Open | 14,686 (4.5%) | 9 (7.8%) | 0.10 |
| External fixation | 7724 (2.4%) | 14 (12.1%) | <0.001 |
| Internal fixation without reduction | 12,544 (3.9%) | 12 (10.3%) | <0.001 |
| Closed reduction without external fixation | 8486 (2.6%) | 14 (12.1%) | <0.001 |
| Closed reduction with external fixation | 49,539 (15.3%) | 30 (25.9%) | 0.002 |
| Open reduction without fixation | 1447 (0.4%) | 0 | 0.47 |
| Open reduction and internal fixation | 92,261 (28.5%) | 46 (39.7%) | 0.01 |
| Debridement | 9141 (2.8%) | 7 (6.0%) | 0.04 |
| All operative | 156,523 (48.3%) | 87 (75.0%) | <0.001 |
| Non-operative | 76,055 (32.9%) | 14 (13.9%) | <0.001 |
| Patella fracture, n (%) | |||
| Patella closed | 5951 (1.8%) | 6 (5.2%) | 0.01 |
| Patella open | 2736 (0.8%) | 2 (1.7%) | 0.30 |
| Tibia fracture, n (%) | |||
| Closed | 56,397 (17.4%) | 26 (22.4%) | 0.16 |
| Open | 34,303 (10.6%) | 19 (16.4%) | 0.04 |
| External fixation | 18,498 (5.7%) | 14 (12.1%) | 0.003 |
| Internal fixation without reduction | 3843 (1.2%) | 2 (1.7%) | 0.59 |
| Closed reduction without external fixation | 1772 (0.5%) | 0 | 0.43 |
| Closed reduction with external fixation | 10,265 (3.2%) | 6 (5.2%) | 0.22 |
| Open reduction without fixation | 2047 (0.6%) | 1 (0.9%) | 0.75 |
| Open reduction and internal fixation | 45,266 (14.0%) | 25 (21.6%) | 0.02 |
| Debridement | 25,552 (7.9%) | 15 (12.9%) | 0.04 |
| All operative | 65,230 (20.1%) | 39 (33.6%) | <0.001 |
| Non-operative | 26,330 (31.3%) | 10 (23.8%) | 0.31 |
| Multiple fractures, n (%) | 62,631 (19.3%) | 51 (44.0%) | <0.001 |
Note: all procedures performed in the operating room.
3.2. Risk of fat embolism syndrome
On multivariable logistic regression, multiple fractures (OR 2.18, CI 1.08–4.42, p = 0.03), closed femur fractures (OR 4.11, CI 2.18–7.76, p < 0.001), and age ≤30 years-old (OR 5.30, CI 3.62–7.75, p < 0.001) were all associated with an increased risk for FES. There was no associated risk of FES with humerus, tibia, or pelvic fractures (Table 4).
Table 4.
Multivariable analysis for risk of fat embolism in blunt trauma patients with extremity fractures.
| Risk factor | OR | CI | p-value |
|---|---|---|---|
| Multiple fractures | 2.18 | 1.08–4.42 | 0.03 |
| Open femur fracture | 1.54 | 0.73–3.24 | 0.26 |
| Closed femur fracture | 4.11 | 2.18–7.76 | <0.001 |
| Open humerus fracture | 1.23 | 0.37–4.08 | 0.73 |
| Closed humerus fracture | 1.00 | 0.48–2.08 | 1.00 |
| Open tibia fracture | 1.76 | 0.95–3.26 | 0.07 |
| Closed tibia fracture | 1.26 | 0.70–2.26 | 0.44 |
| Age ≤30 years | 5.30 | 3.62–7.75 | <0.001 |
| Pelvic fracture | 0.76 | 0.40–1.46 | 0.41 |
In blunt trauma patients with femur fractures, age ≤30 (OR 5.13, CI 3.41–7.71, p < 0.001) and multiple fractures (OR 2.14, CI 1.43–3.20, p < 0.001) were both independently associated with increased risk of FES. Fractures amenable to non-surgical management were associated with a decreased risk of FES (OR 0.44, CI 0.25–0.77, p = 0.004) (Table 5).
Table 5.
Multivariable analysis for risk of fat embolism in blunt trauma patients with femur fractures.
| Risk factor | OR | CI | p-value |
|---|---|---|---|
| Age ≤30 years | 5.13 | 3.41–7.71 | <0.001 |
| Non-operative | 0.44 | 0.25–0.77 | 0.004 |
| Multiple fractures | 2.14 | 1.43–3.20 | <0.001 |
There was no association with FES for non-operative treatment as compared to operative treatment in tibia or humerus fractures (p > 0.05). Multiple fractures (OR 16.30, CI 2.13–124.6, p = 0.01) as well as age ≤30 years-old (OR 8.20, CI 2.53–26.56, p < 0.001) were associated with an increased risk of FES in patients with humerus fractures. In patients with tibia fractures, multiple fractures (OR 6.71, CI 3.10–14.53, p < 0.001) and age ≤30 (OR 3.76, CI 2.03–6.97, p < 0.001) were associated with an increased risk of FES (Table 6, Table 7).
Table 6.
Multivariable analysis for risk of fat embolism in blunt trauma patients with humerus fractures.
| Risk factor | OR | CI | p-value |
|---|---|---|---|
| Age ≤30 years | 8.20 | 2.53–26.56 | <0.001 |
| Non-operative | 0.83 | 0.28–2.52 | 0.75 |
| Multiple fractures | 16.30 | 2.13–124.6 | 0.01 |
Table 7.
Multivariable analysis for risk of fat embolism in blunt trauma patients with tibia fractures.
| Risk factor | OR | CI | p-value |
|---|---|---|---|
| Age ≤30 years | 3.76 | 2.03–6.97 | <0.001 |
| Non-operative | 0.58 | 0.28–1.18 | 0.13 |
| Multiple fractures | 6.71 | 3.10–14.53 | <0.001 |
In blunt trauma patients with extremity fractures and other serious regional body injuries, defined as a grade >2 for abbreviated injury scale (AIS), lower extremity injuries (OR 7.59, CI 2.73–21.06, p < 0.001) and thoracic injuries (OR 2.76, CI 1.76–4.32, p < 0.001) were associated with an increased risk of FES. There was no associated risk of FES for patients with serious head, abdominal, spinal, or upper extremity injuries (p > 0.05) (Table 8).
Table 8.
Multivariable analysis for risk of fat embolism from serious body region injuries (abbreviated injury scale grade >2) in blunt trauma patients.
| Risk factor | OR | CI | p-value |
|---|---|---|---|
| Head | 1.13 | 0.66–1.96 | 0.65 |
| Thorax | 2.76 | 1.76–4.32 | <0.001 |
| Abdomen | 1.30 | 0.68–2.49 | 0.43 |
| Spine | 0.93 | 0.37–2.32 | 0.87 |
| Lower extremity | 7.59 | 2.73–21.06 | <0.001 |
| Upper extremity | 1.32 | 0.65–2.67 | 0.45 |
3.3. Hospital outcomes
Compared to blunt trauma patients without FES, the FES patients had an increased mortality rate (11.2% vs. 4.1%, p < 0.001). The FES group also had an increased median hospital LOS (12 vs. 6 days, p < 0.001), ICU LOS (6 vs. 4, p = 0.02), and complications of ARDS (22.4% vs. 1.2%, p < 0.001), extremity compartment syndrome (2.6% vs. 0.5%, p = 0.002), CVA (5.2% vs. 0.4%, p < 0.001), unplanned intubations (10.3% vs. 1.2%, p < 0.001), PE (7.8% vs. 0.9%, p < 0.001), and pneumonia (18.1% vs. 3.3%, p < 0.001) (Table 9).
Table 9.
Outcomes of blunt trauma patients with extremity fractures.
| Without embolism | Fat embolism | ||
|---|---|---|---|
| Characteristic | (n = 324,049) | (n = 116) | p-value |
| LOS, days, median (IQR) | 6.0 (6, 12) | 12.0 (9, 22) | <0.001 |
| ICU, days, median (IQR) | 4.0 (4, 11) | 6.0 (4, 13) | 0.02 |
| Ventilator, days, median (IQR) | 4.0 (4, 12) | 7.0 (5, 14) | 0.30 |
| Complications, n (%) | |||
| Acute kidney injury | 4417 (1.4%) | 4 (3.4%) | 0.05 |
| ARDS | 4013 (1.2%) | 26 (22.4%) | <0.001 |
| Cardiac arrest with CPR | 3709 (1.1%) | 5 (4.3%) | 0.001 |
| CAUTI | 120 (0.04%) | 0 | 0.84 |
| CLASBI | 15 (0.005%) | 0 | 0.94 |
| Decubitus | 3505 (1.1%) | 3 (2.6%) | 0.12 |
| Deep site infection | 990 (0.3%) | 0 | 0.55 |
| Extremity compartment syndrome | 1713 (0.5%) | 3 (2.6%) | 0.002 |
| Myocardial infarction | 1842 (0.6%) | 0 | 0.42 |
| Organ space infection | 893 (0.3%) | 0 | 0.57 |
| Cerebrovascular accident | 1389 (0.4%) | 6 (5.2%) | <0.001 |
| Superficial infection | 1151 (0.4%) | 1 (0.9%) | 0.36 |
| Unplanned ICU admission | 3725 (1.1%) | 5 (4.3%) | 0.001 |
| Unplanned intubation | 3976 (1.2%) | 12 (10.3%) | <0.001 |
| Deep vein thrombosis | 5617 (1.7%) | 9 (7.8%) | <0.001 |
| Pulmonary embolism | 2901 (0.9%) | 9 (7.8%) | <0.001 |
| Pneumonia | 10,767 (3.3%) | 21 (18.1%) | <0.001 |
| Osteomyelitis | 208 (0.1%) | 0 | 0.79 |
| Severe sepsis | 1840 (0.6%) | 0 | 0.42 |
| Urinary tract infection | 8605 (2.7%) | 5 (4.3%) | 0.27 |
| Ventilator associated pneumonia | 148 (0.05%) | 0 | 0.82 |
| Mortality, n (%) | 13,377 (4.1%) | 13 (11.2%) | <0.001 |
IQR = interquartile range, LOS = length of stay, ICU = intensive care unit, ARDS = acute respiratory distress syndrome, CAUTI = catheter associated urinary tract infection, CLABSI = central line associated bloodstream infection, CPR = cardiopulmonary resuscitation.
4. Discussion
FES is a relatively rare although potentially fatal complication following long bone fractures in high energy trauma. To our knowledge, this is the first study in the United States to analyze and evaluate outcomes in FES patients using the TQIP large, national database. In our analysis spanning seven years, the rate of FES in blunt trauma patients with long-bone extremity fractures was less than 0.1%. Age ≤30 years-old, closed femur fractures, and multiple long bone fractures were all associated with an increased risk of FES. In patients with femur fractures, fractures amenable to non-surgical management were associated with a decreased risk of FES. There was no association for operative versus non-operative treatment in humerus fractures or tibia fractures. There was no increased risk of FES in patients with pelvic fractures.
Patients with serious thoracic injuries or serious lower extremity injuries (AIS grade >2) had an increased associated risk of developing FES. Future evaluation of this finding with serious thoracic trauma is needed to determine if this is merely an association or there may be causation with fat embolism from chest wall fractures.
Fractures have long been known to be associated with FES in trauma patients. Bulger et al. conducted a retrospective 10-year review at a single Level I trauma center to determine the demographics, injury severity and pattern, diagnostic criteria, and management of all patients diagnosed with FES. They reported that during this time, 27 blunt trauma patients were diagnosed with FES, resulting in an incidence of almost 1% of all patients with long-bone fractures. There were a total of 40 long-bone fractures in the patients, 22 (55%) of which were femur fractures and 10 (37%) of the patients had associated pelvic fractures.25 Robert et al. over a 25-year study period found a 0.26% (20 patients) incidence of FES in trauma patients with 70% of patients suffering from femur fractures, 80% from tibia fractures, and 15% with pelvic fractures.35 In support of the high rate of FES after a femur fracture, we specifically found that closed femur fractures were associated with an over four-times increased risk of FES when compared to open femur fractures. On multivariable analysis, however, pelvic fractures and tibia fractures were not independent predictors for FES. This is in contrast to previous reports that pelvic fractures are associated with FES.21,41 This may be because most of the marrow in the flat bones of the adult skeleton including the pelvic bones is comprised of red bone marrow which has a lower fat content compared to yellow bone marrow (which is more prominent in the long bones).23 Although clinicians are unable to change a patient's risk factors for fat embolism, having the knowledge that closed femur fractures increase the risk of developing FES is of paramount importance so that a high index of clinical suspicion for these patients can be maintained, ultimately leading to earlier diagnosis and treatment.26
The occurrence of multiple fractures, each of which may put the patient at risk for FES, has previously been demonstrated to be associated with an increased rate of FES.37,38 Stein et al. looked at the incidence of FES in over 900,000,000 patients getting discharged from short-stay hospitals in the United States. They reported that among patients with multiple fractures including the femur (excluding neck), over 1% of them had FES in comparison to less than 1% of patients with isolated femur fractures (excluding neck).39 Tsai et al. performed a 12-year retrospective study on patients with long bone fractures in a tertiary referral center and found that the incidence of FES in patients with multiple fractures was over 2% versus less than 1% in an isolated femur fracture and 0.15% in an isolated tibia fracture. We add to this literature by demonstrating that when controlling for age and other significant covariates (e.g. fracture types), that multiple fractures was independently associated with an over two-fold increased risk of FES. Therefore, vigilance for FES should be maintained when treating patients with multiple fractures.
Interestingly, we found that age ≤30 to be the strongest predictor of FES with an over five-fold associated increased risk of FES. In support of this, Tsai et al. found that in their 13 patients who developed FES, all except one were less than 35 years-old.9 Similarly, Stein et al. found that the patients who developed FES were more often between the ages of 10 and 39 years-old.39 One possible explanation for the more frequent occurrence of FES in younger patients is the fracture location within the long bone. In older patients, low-energy trauma often results in typical geriatric fracture patterns such as femoral neck and intertrochanteric fractures. These patterns occur in a location where there is no fat-containing medullary canal.28 This results in a lower chance for the development of FES. Future research to better elucidate if age or fracture location is the true risk factor is needed. Additionally, incorporation of our findings into a risk-stratification tool may help identify the patients at most risk for FES and thereby lead to more prompt diagnosis and/or studies regarding targeted interventions within a high-risk population.
There have been previously conflicting results regarding the association of FES with fracture surgery.32,33,36 The majority of femoral shaft fractures are treated operatively with a medullary nail. In a prospective, consecutive, nonrandomized clinical trial, Kropfl et al. looked at 36 adult patients with femur fractures treated surgically with femoral nailing, separated into reamed and unreamed cohorts. They measured the intramedullary pressure during the operation and the bone marrow fat intravasation. The reamed nailing cohort had a significantly higher intramedullary pressure (396 ± 85 mmHg) and higher bone marrow fat intravasation (3.2 ± 0.4), compared to the unreamed nailing cohort, due to the fat release during the reaming process (intramedullary pressure 91 ± 26 mm Hg, fat intravasation 1.9 ± 0.2).40 Increased intramedullary pressure, classically thought to arise from reaming and nail insertion during intramedullary nailing of fractures, can theoretically lead to an embolization of fat resulting in FES.20,34 In addition to the method of fracture surgery, the timing of the procedure is another important consideration. Early stabilization of fractures may lead to a decreased rate of FES.30,31 In our cohort, femur fractures amenable to non-operative treatment were associated with a decreased risk of FES, though this may be related to the effect of fracture pattern not captured by ICD-9 diagnosis. This was not the case for humerus fractures or tibia fractures, potentially because of the different fixation techniques used for the different long bones. Knowing that surgically treated femur fractures are associated with a higher risk of FES is important so clinicians can remain vigilant during these procedures and minimize increases in intramedullary pressure. In addition, future combined basic science and clinical research to better clarify what role operative technique has in the development of FES appears warranted.
There are several limitations to our study including the possibility of misclassification and missing data from the use of a large retrospective database. Also the TQIP database only includes trauma patients in trauma centers, and thus only includes patients who have significant mechanisms and/or injuries and thus does not incorporate all patients presenting to a hospital with extremity fractures. Additionally, the diagnosis for FES was not standardized in the dataset nor were the findings related to FES diagnosis and/or complications such as neurologic dysfunction, respiratory failure and petechial rash reported. In addition, missing pertinent data variables include the time from injury to diagnosis of FES, type and location of fracture(s) (e.g. proximal, distal, neck etc.), indication for operative fixation, details regarding exact operative techniques used for surgical stabilization of fractures (e.g. intramedullary instrumentation, nailing, plating etc.), the length of time from injury to operation, the timing of fracture stabilization, and the exact definitions of the TQIP terminology used for the procedures. Lastly, TQIP is limited to the index hospitalization, thus does not provide post-discharge information regarding complications, mortality, or long-term functional outcomes. Additionally, any post-discharge cases of FES would not have been captured. Despite these limitations our study is strengthened by its large nationwide sample and robust multivariable analysis controlling for known predictors of FES.
In summary, this retrospective national analysis found the incidence of FES in blunt trauma patients sustaining long bone extremity fractures to be less than 0.1%, however the mortality rate for FES was over 11%. In addition, we identified closed femur fractures, age ≤30 years-old, and multiple fractures to be associated with increased risk for FES. Future research is warranted to validate these results and use these findings to help develop a clinical risk stratification tool to identify patients at highest risk for FES with the hopes of reducing the significant associated morbidity and mortality.
Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Declaration of competing interest
Dr. James Learned reports personal fees from DePuy Synthes, personal fees from Abyryx, personal fees from Smith and Nephew, outside the submitted work. None of the other authors have anything to disclose.
Miriam Alpert, Dr. Areg Grigorian, and Dr. Jeffry Nahmias researched literature, conceived the study, and were involved in data analysis. Miriam Alpert wrote the first draft of the manuscript. All authors helped with the study design, data interpretation and reviewed and edited the manuscript.
Contributor Information
Miriam Alpert, Email: Miriam.alpert@westernu.edu.
Areg Grigorian, Email: agrigori@uci.edu.
John Scolaro, Email: jscolaro@uci.edu.
James Learned, Email: jlearned@uci.edu.
Matthew Dolich, Email: mdolich@hs.uci.edu.
Catherine M. Kuza, Email: catherine.kuza@gmail.com.
Michael Lekawa, Email: melekawa@uci.edu.
Jeffry Nahmias, Email: jnahmias@uci.edu.
References
- 1.Kwiatt M.E., Seamon M.J. Fat embolism syndrome. Int J Crit Illn Inj Sci. 2013;3(1):64–68. doi: 10.4103/2229-5151.109426. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Weinhouse Gerald L. Fat Embolism Syndrome. Up to Date. March 2019. https://www.uptodate.com/contents/fat-embolism-syndrome?search=fat%20embolism&source=search_result&selectedTitle=1~72&usage_type=default&display_rank=1
- 3.Mellor A., Soni N. Fat embolism. Anaesthesia. 2001;56(2):145–154. doi: 10.1046/j.1365-2044.2001.01724.x. [DOI] [PubMed] [Google Scholar]
- 4.Fabian T., Hoots A., Stanford D., Patterson C., Mangiante E. Fat-embolism syndrome - prospective evaluation in 92 fracture patients. Crit Care Med. 1990;18(1):42–46. https://www.ncbi.nlm.nih.gov/pubmed/2293968 [PubMed] [Google Scholar]
- 5.Duis H.J., Nijsten M.W., Klasen H.J., Binnendijk B. Fat embolism in patients with an isolated fracture of the femoral shaft. J Trauma. 1988;28:383–390. doi: 10.1097/00005373-198803000-00015. [DOI] [PubMed] [Google Scholar]
- 6.Cronin K.J., Hayes C.B., Moghadamian E.S. Early-onset fat embolism syndrome. Case Connector. 2018;8(2):1–5. doi: 10.2106/JBJS.CC.17.00175. [DOI] [PubMed] [Google Scholar]
- 7.Fukumoto L.E., Fukumoto K.D. Fat embolism syndrome. Nurs Clin. 2018;53(3):335–+. doi: 10.1016/j.cnur.2018.04.003. [DOI] [PubMed] [Google Scholar]
- 8.Huang B.K., Ju Monu, Wandtke J., Huang B.K., Monu J.U.V., Wandtke J. Pulmonary fat embolism after pelvic and long bone fractures in a trauma patient. Emerg Radiol. 2009;16(5):407–409. doi: 10.1007/s10140-008-0757-7. [DOI] [PubMed] [Google Scholar]
- 9.Agustin Godoy D., Di Napoli M., Rabinstein A.A. Cerebral fat embolism: recognition, complications, and prognosis. Neurocritical Care. 2018;29(3):358–365. doi: 10.1007/s12028-017-0463-y. [DOI] [PubMed] [Google Scholar]
- 10.Tzioupis C.C., Giannoudis P.V. Fat embolism syndrome: what have we learned over the years? Trauma. 2011;13:259–281. [Google Scholar]
- 11.Özgül Ülkü, Gedik Ender, Karakaplan Mustafa. Fat embolism syndrome in two cases with multiple fractures. Türk Yoğun Bakim Derneği Dergisi. 2012;1:23. doi: 10.4274/Tybdd.146. [DOI] [Google Scholar]
- 12.Husebye E.E., Lyberg T., Roise O. Bone marrow fat in the circulation: clinical entities and pathophysiological mechanisms. Injury. 2006;37(suppl 4):S8–S18. doi: 10.1016/j.injury.2006.08.036. [DOI] [PubMed] [Google Scholar]
- 13.Karagiorga G., Nakos G., Galiatsou E., Lekka M.E. Biochemical parameters of bronchoalveolar lavage fluid in fat embolism. Intensive Care Med. 2006;32:116–123. doi: 10.1007/s00134-005-2868-x. [DOI] [PubMed] [Google Scholar]
- 14.Gurd A.R., Wilson R.I. The fat embolism syndrome. The Journal Of Bone And Joint Surgery British. 1974;56B(3):408–416. https://www.ncbi.nlm.nih.gov/pubmed/4547466 [PubMed] [Google Scholar]
- 15.Wong M., Tsui H., Yung S., Chan K., Cheng J. Continuous pulse oximeter monitoring for inapparent hypoxemia after long bone fractures. J Trauma. 2004;56(2):356–362. doi: 10.1097/01.TA.0000064450.02273.9B. [DOI] [PubMed] [Google Scholar]
- 16.Tsai I.-T., Hsu C.-J., Chen Y.-H., Fong Y.-C., Hsu H.-C., Tsai C.-H. Fat embolism syndrome in long bone fracture—clinical experience in a tertiary referral center in Taiwan. J Chin Med Assoc. 2010;73(8):407–410. doi: 10.1016/S1726-4901(10)70088-5. [DOI] [PubMed] [Google Scholar]
- 17.Shareef K.A., Asadullah M., Helal M. Fat embolism syndrome due to fracture right femur: a case report. Egyptian Journal of Hospital Medicine. 2017;68(1):923–928. doi: 10.12816/0038192. [DOI] [Google Scholar]
- 18.Bone L., Johnson K., Weigelt J., Scheinberg R. Early versus delayed stabilization of femoral fractures - a prospective randomized study. J Bone Joint Surg Am. 1989;71A(3):336–340. https://www.ncbi.nlm.nih.gov/pubmed/2925704 [PubMed] [Google Scholar]
- 19.Riska E., Myllynen P. Fat embolism in patients with multiple injuries. Orthop Trauma Dir. 2009;7(6):29–33. doi: 10.1055/s-0028-1100871. [DOI] [PubMed] [Google Scholar]
- 20.Habashi N.M., Andrews P.L., Scalea T.M. Therapeutic aspects of fat embolism syndrome. Injury. 2006;37(4):S68–S73. doi: 10.1016/j.injury.2006.08.042. [DOI] [PubMed] [Google Scholar]
- 21.Koul Parvaiz A., Ahmad Feroze, Gurcoo Showkat A. Fat embolism syndrome in long bone trauma following vehicular accidents: experience from a tertiary care hospital in north India. Lung India. 2013;(2):97. doi: 10.4103/0970-2113.110413. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Trauma Quality Improvement Program . 2017. American College of Surgeons.https://www.facs.org/quality-programs/trauma/tqp/center-programs/tqip [Google Scholar]
- 23.Małkiewicz A., Dziedzic M. Bone marrow reconversion - imaging of physiological changes in bone marrow. Pol J Radiol. 2012;77(4):45–50. doi: 10.12659/pjr.883628. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Aggarwal Richa, Banerjee Arnab, Kapil dev Soni. Kumar Atin, Trikha Anjan. Clinical characteristics and management of patients with fat embolism syndrome in level I Apex Trauma Centre. Chin J Traumatol. 2019;3:172. doi: 10.1016/j.cjtee.2019.01.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Bulger E., Smith D., Maier R., Jurkovich G. Fat embolism syndrome - a 10-year review. Arch Surg. 1997;132(4):435–439. doi: 10.1001/archsurg.1997.01430280109019. https://www.ncbi.nlm.nih.gov/pubmed/9108767 [DOI] [PubMed] [Google Scholar]
- 26.George J., George R., Dixit R., Gupta R.C., Gupta N. Fat embolism syndrome [published correction appears in Lung India. Lung India. 2013;34(1):47–53. doi: 10.4103/0970-2113.106133. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Berlot G., Bussani R., Shafiei V., Zarrillo N. July 2018. Fulminant Cerebral Fat Embolism: Case Description and Review of the Literature; pp. 1–5. Case Reports in Critical Care. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Dillerud E. Abdominoplasty combined with suction lipoplasty - a study of complications, revisions, and risk-factors in 487 cases. Ann Plast Surg. 1990;25(5):333–343. doi: 10.1097/00000637-199011000-00001. https://www.ncbi.nlm.nih.gov/pubmed/2147821 [DOI] [PubMed] [Google Scholar]
- 29.Shah F.A., Alam W., Khan W.M. Fat embolism syndrome in long bone fractures: our experience at Lady Reading Hospital Peshawar. Pakistan Journal of Surgery. 2017;33(1):25–29. http://www.pjs.com.pk/journal_pdfs/jan_mar17/25.pdf [Google Scholar]
- 30.Rossi S., Goodman P., Franquet T. Nonthrombotic pulmonary emboli. AJR Am J Roentgenol. 2000;174(6):1499–1508. doi: 10.2214/ajr.174.6.1741499. https://www.ncbi.nlm.nih.gov/pubmed/10845470 [DOI] [PubMed] [Google Scholar]
- 31.Robinson C. Current concepts of respiratory insufficiency syndromes after fracture. J Bone Joint Surg Br. 2001;83B(6):781–791. doi: 10.1302/0301-620x.83b6.12440. https://www.ncbi.nlm.nih.gov/pubmed/11521914 [DOI] [PubMed] [Google Scholar]
- 32.Janio Jose Alves Bezerra Silva, Diogo de Almeida Diana, Victor Eduardo Roman Salas. Zamboni Caio, Jose Soares Hungria Neto. Christian Ralph Walter. Fat embolism syndrome in femoral shaft fractures: does the initial treatment make a difference? Rev Bras Ortop. 2017;(5):535. doi: 10.1016/j.rboe.2016.08.021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Aggarwal R., Pal S., Soni K.D., Gamangatti S. Massive cerebral fat embolism leading to brain death: a rare presentation. Indian J Crit Care Med. 2015;19(11):687–689. doi: 10.4103/0972-5229.169358. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Nissar Shaikh. Emergency management of fat embolism syndrome. J Emergencies, Trauma, Shock. 2009;1:29. doi: 10.4103/0974-2700.44680. https://www.ncbi.nlm.nih.gov/pubmed/19561953 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Lee S.C., Yoon J.Y., Nam C.H., Kim T.K., Jung K.A., Lee D.W. Cerebral fat embolism syndrome after simultaneous bilateral total knee arthroplasty: a case series. J Arthroplasty. 2012;27(3):409–414. doi: 10.1016/j.arth.2011.06.013. [DOI] [PubMed] [Google Scholar]
- 36.Randelli F., Capitani P., Pace F., Favilla S., Galante C., Randelli P. Bilateral femoral shaft fractures complicated by fat and pulmonary embolism: a case report. Injury. 2015;46:S28–S30. doi: 10.1016/S0020-1383(15)30041-3. https://www.ncbi.nlm.nih.gov/pubmed/26738456 [DOI] [PubMed] [Google Scholar]
- 37.Gore T., Lacey S. Bone up on fat embolism syndrome. Nursing. 2005;35(8):32hn1–32hn4. doi: 10.1097/00152193-200508000-00026. [DOI] [Google Scholar]
- 38.Scarpino M., Lanzo G., Lolli F., Grippo A. From the diagnosis to the therapeutic management: cerebral fat embolism, a clinical challenge. Int J Gen Med. 2019;39 doi: 10.2147/IJGM.S177407. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6324602/ [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Stein P.D., Yaekoub Mayf, Kleerekoper M. Fat embolism syndrome. Am J Med Sci. 2008;336(6):472–477. doi: 10.1097/MAJ.0b013e318172f5d2. [DOI] [PubMed] [Google Scholar]
- 40.Kropfl A., Berger U., Neureiter H., Hertz H., Schlag G. Intramedullary pressure and bone marrow fat intravasation in unreamed femoral nailing. J Trauma. 1997;42(5):946–954. doi: 10.1097/00005373-199705000-00028. https://www.ncbi.nlm.nih.gov/pubmed/9191679 [DOI] [PubMed] [Google Scholar]
- 41.Gupta B., D’souza N., Sawhney C. Analyzing fat embolism syndrome in trauma patients at AIIMS Apex Trauma Center, New Delhi, India. J Emergencies, Trauma, Shock. 2011;4(3):337–341. doi: 10.4103/0974-2700.83859. [DOI] [PMC free article] [PubMed] [Google Scholar]