Single-case metanalysis of fat embolism syndrome

Single-case metanalysis of fat embolism syndromeInternational Journal of Cardiology 345 (2021) 111–117
Available online 30 October 2021 0167-5273/© 2021 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Single-case metanalysis of fat embolism syndrome
Zuowen He a,b, Zeqi Shi b, Chenze Li b, Li Ni a,b, Yang Sun a,b, Francesco Arioli c, Yan Wang a,b, Enrico Ammirati d,**,1, Dao Wen Wang a,b,*,1
a Division of Cardiology, Department of Internal Medicine, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China b Hubei Key Laboratory of Genetics and Molecular Mechanism of Cardiological Disorders, Wuhan, China c Department of Cardiology, Fatebenefratelli Hospital, Milano, Italy d De Gasperis Cardio Center and Transplant Center, Niguarda Hospital, Milano, Italy
A R T I C L E I N F O
Keywords: Fat embolism syndrome Pulmonary embolism Respiratory failure Bone fracture Corticosteroids
A B S T R A C T
Background: Only one large series has been reported on fat embolism syndrome (FES), a condition caused by fat globules release into the circulation, primarily as consequence of bone fracture. Thus, more data on clinical features, therapies, and prognosis are needed. Methods and results: The study screened 1090 manuscripts in PubMed and Web of Science on cases of FES published from June 2010 to June 2020. The authors identified 124 studies and included in the pooled-analysis 135 patients (>14 years), plus one additional unpublished case managed in Tongji hospital. All had confirmed diagnosis of FES with complete clinical data. The median age at presentation was 39 years, and 82 (61.8%) were men. FES was predominantly associated with bone fractures (78, 57.4%), particularly femur fracture (59, 43.4%). The most common clinical finding at the onset was respiratory abnormalities in 34.6% of all clinical presentations. Therapies included respiratory supportive care in 127 (93.4%) patients, application of cortico- steroids in 22 (16.2%) and anticoagulant in 5 (3.7%) cases. Overall mortality was 30.2% (N = 41), and logistic regression analysis showed that corticosteroid therapy was significantly associated with reduced mortality with an OR of 0.143 (95%CI 0.029–0.711), while age ≥ 65 years and non-orthopedic conditions were significantly associated with increased mortality with an OR of 4.816 (95%CI 1.638–14.160) and 4.785 (95%CI 1.019–22.474). Conclusions: FES has been associated with a larger mortality rate than previously observed, although publication bias can have led to overestimation of mortality. Finally, a potential protective effect of corticosteroid therapy has been suggested by the current analysis.
1. Introduction
Fat embolism syndrome (FES) is a rare clinical syndrome that follows an identifiable injury [1], which releases fat globules into the circula- tion, resulting in pulmonary and systemic signs and symptoms [2]. FES typically manifests 24–72 h after the initial event, and usually follows bone fractures or orthopedic operations [3]. Patients often have an insidious clinical onset with respiratory abnormalities (i.e. dyspnea, hypoxemia and tachypnea), neurologic abnormalities and less frequently petechial rash [4]. Furthermore, sudden cardiopulmonary collapse can occur as a consequence of massive fat embolism [5].
Diagnosis can be challenging, as no specific diagnostic tools are avail- able. Identification of patients with FES relies on known causal triggers, clinical signs and symptoms, laboratory results, imaging findings in the appropriate clinical setting, and exclusion of other differential diagnosis [3]. FES is a rare clinical syndrome, with a poor understanding of related pathophysiology. The mortality rate is 6–10% [1,6], often associated with a lag in the diagnosis. Therefore, FES awareness and an accurate description of clinical presentation and associated conditions are essential for its early recognition. To the best of our knowledge, only one large registry has analyzed the clinical characteristics of patients with FES limited to a specific subpopulation of patients with FES after trauma
* Correspondence to: D. W. Wang, Division of Cardiology, Departments of Internal Medicine, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, 1095# Jiefang Ave, Wuhan 430030, China. ** Correspondence to: E. Ammirati, De Gasperis Cardio Center and Transplant Center, Niguarda Hospital, Piazza Ospedale Maggiore 3, 20162 Milano, Italy.
E-mail addresses: [email protected] (E. Ammirati), [email protected] (D.W. Wang). 1 E.A and D.W.W. share senior authorship.
Contents lists available at ScienceDirect
International Journal of Cardiology
112
[1]. We performed a pooled analysis of 135 such cases of FES due to any cause retrieved from 124 isolated reports plus an additional unpublished case. The primary objectives were to investigate clinical features, ther- apies, and prognosis of FES.
2. Methods
2.1.1. Study design This pooled-analysis was performed according to the Preferred
Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. [7] The study protocol was registered with the international prospective register of systematic reviews (PROSPERO) (CRD42020220017).
Search strategy: The PubMed database was searched (Search param- eters: time from June 25, 2010, to June 25, 2020, English language, human species) for all studies describing FES (keywords: “fat embolism syndrome”), and the Web of Science was searched (Search parameters: time from June 25, 2010, to June 25, 2020, English language) for all studies describing FES (keywords: “fat embolism syndrome”).
2.2. Eligibility criteria
We included all studies in which: (1) diagnosis of FES was confirmed, (2) patients were older than 14 years, (3) available clinical data were complete, including etiologies, clinical presentations, diagnosis pat- terns, auxiliary examinations, treatments and in-hospital outcome (Supplementary Fig. 1). Among the considered cases we included an unpublished case of a 31-year-old woman diagnosed with FES at the Tongji Hospital (see Fig. 2 for a brief presentation of the case and Supplemental Materials for the full presentation of the case).
2.3. Data extraction
A standardized data extraction database was created by tabulating the following data: sex, age, comorbidity, associated conditions, fracture sites, clinical presentations, auxiliary examination performance, diag- nostic mode, treatments and prognosis. Z.H and Z.S conducted the literature search independently to verify data accuracy and complete- ness, with a third reviewer resolving any uncertainties.
2.4. Term definitions
When we identified the associated conditions, we termed “orthope- dic operations” interventions such as total hip replacement, knee arthroplasty, internal fixation and spinal instrumentation. “Non-ortho- pedic conditions” include liposuction or autologous fat injection, sickle cell disease, exogenous fat-soluble substances injection, renal angio- myolipoma, lung transplantation, coronary artery bypass graft, hepa- tocellular carcinoma, multiple soft tissue injuries and hypertriglyceridemia.
Regarding clinical presentation, respiratory abnormalities are defined as the presence of at least one of the following presentations: tachypnea, hypoxemia, dyspnea, chest pain, rale, hemoptysis, and cyanosis. Neurological abnormalities are defined as the presence of at least one of the following presentations: disturbance of consciousness, restlessness, seizures, limb weakness, paralysis, speech disorder, sensory disturbance, headache and dizziness. Cardiovascular abnormalities are defined as the presence of at least one of the following presentations: tachycardia, hypotension, cardiac arrest, bradycardia and angina.
The term neurological therapies/interventions included sedatives, and/or anticonvulsant drugs, and/or interventional thrombectomy.
Regarding echocardiography, right ventricular abnormality was defined as right ventricular hypokinesis, and/or dilation, and/or dysfunction; left ventricular abnormality was defined as left ventricular
hypokinesis, and/or dysfunction.
2.5. Statistical analysis
Categorical variables were described as percentages and frequency rates, and continuous variables were described using median and first to third quartile (Q1-Q3) values. Median for continuous variables were compared using the Mann-Whitney U test. Proportions for categorical variables were compared using the χ2 test. Univariate logistic regression analysis was used to identify the effect of baseline, associated condi- tions, comorbidity and treatment modes on mortality. Furthermore, multivariable logistic regression was performed to analyze variables with p < 0.10 in univariate logistic regression on mortality. All of the statistical analyses were performed using the Statistical Package for the Social Sciences version 17.0 software. For unadjusted comparisons, a 2- sided α of <0.05 was considered to be significant.
3. Results
3.1. Selection process and general information
The initial search in PubMed for “fat embolism syndrome” with limitation of “human species” and “English languages” yielded 560 studies from June 25, 2010, to June 25, 2020. The initial search in Web of Science for “fat embolism syndrome” with limitation of “English language” yielded 530 studies from June 25, 2010, to June 25, 2020. A total of 1090 studies were identified from the two databases (PubMed and Web of science). Of the 1090 studies, a total of 966 studies were excluded, which included 533 meeting abstracts, unrelated studies, clinical studies, books, editorial materials, letters, corrections, reviews or experimental studies, 381 duplicates, 44 studies with incomplete data, 4 studies with patients ≤14 years and 4 studies not published in English. Finally, 124 articles containing 135 patients with FES plus the additional unpublished case from our hospital were enrolled for anal- ysis. Data on the 136 individual cases were merged and analyzed (Supplementary Fig. 1). FES was mostly reported from hospitals in North America (38.1%), Europe (29.1%), and Asia (24.6%). The remaining cases were reported from hospitals in South America (3.7%), New Zealand and Australia (3.7%), and Africa (0.7%) (Supplemental Fig. 2A). The full list of the 124 articles is available in the Supplemental Materials.
The median age at presentation was 39 years (Q1-Q3: 24–63; range: 15–93 years), and 84 (61.8%) were men. Fifty-three patients (39.0%) had one or more coexisting medical condition. Among the common pre- existing conditions, 11 patients had hypertension (8.1%), 5 had diabetes (3.7%), 4 (2.9%) had arrhythmia, 4 (2.9%) had cancer, and 4 (2.9%) had a connective tissue disease. Of these 136 patients, 41 (30.2%) died (Table 1).
Compared with survivors (n = 95), non-survivors (n = 41) were significantly older (median age, 55 years [Q1-Q3: 33–78] vs. 33 years [Q1-Q3, 22–56]; p < 0.001) (Table 1). The proportion of non-survivors with age equal or above 65 years was significantly larger compared with survivors, comprising 17/41 (41.5%) vs. 16/95 (16.8%; p = 0.002), respectively (Table 1). Non-survivors had more pre-existing comorbid- ities compared with survivors, 22/41 (53.7%) vs. 31/95 (32.6%; p = 0.021), respectively (Table 1).
3.2. Associated conditions
FES was more frequently associated with bone fractures in 78 (57.4%) patients, orthopedic operations in 67 (49.3%) patients, lipo- suction or autologous fat injection in 16 (11.8%) patients, and sickle cell disease in 14 (10.3%) patients. Other infrequent conditions associated with FES are also reported in Table 1 and Supplementary Fig. 2B. Of 98 (72.1%) cases where the time from the associated condition to the occurrence of FES had been reported, in 61 (62.2%) FES occurred within
Z. He et al.
113
24 h after exposure to the causative factor, in 35 (35.7%) FES occurred between 24 and 48 h later, and in 2 (2.0%) it occurred beyond 48 h (Supplementary Fig. 2C). The mean time from initial (or “associated”) condition to occurrence of FES was 23 h. Furthermore, FES was more commonly associated with femur fracture, reported in 59 (43.4%) pa- tients, tibiofibular fracture, in 25 (18.4%) patients, and spine fracture, in 8 (5.9%) patients. Other infrequent fracture sites are also shown in Supplementary Fig. 2D. There were only 21 (15.4%) cases where fracture type had been reported (closed vs. open fracture). Twelve (57.1%) patients had closed fractures, 7 (33.3%) patients had open fractures and 2 (9.5%) patients had both (Supplementary Fig. 2E).
3.3. Clinical presentation
The most common clinical finding at the onset was respiratory ab- normalities constituting 34.6% of all clinical presentations (Fig. 1A). Specifically, hypoxemia, dyspnea and tachypnea were present in 67 (49.3%), 57 (41.9%) and 26 (19.1%) patients, respectively (Fig. 1B). Neurologic abnormalities were the second most frequent clinical finding constituting 27.3% of all clinical presentations (Fig. 1A). In particular, disturbance of consciousness was observed in 83 (61.0%) patients, and represented the single most common clinical manifestation among pa- tients with FES (Fig. 1B). Cardiovascular manifestations were less frequent, reported in 18.0% of all clinical presentations (Fig. 1A), with tachycardia reported in 40 (29.4%) patients, hypotension in 27 (19.9%) and cardiac arrest in 14 (10.3%) (Fig. 1B). Cardiac arrest was strongly associated with mortality, as it occurred in 13/41 non-survivors vs. 1/95 survivors, p < 0.001 (Supplementary Table 1). Finally, fever and petechial rash were observed in 36 (26.5%) and 33 (24.3%) patients, respectively, while other less common manifestations are also summa- rized in Fig. 1B and Supplementary Table 1.
Table 1 Demographic characteristics, pre-existing conditions, associated conditions, diagnostic evaluation and therapies in the overall population, survivors and non- survivors with FES.
Characteristics All (136) Survivor (95)
Non- survivor
(41)
P-value
Age
Age years, median (Q1-Q3) 39 (24–63) 33 (22–56) 55 (33–78) <0.001
≥65 years 33 (24.3%)
Sex
Associated conditions
Orthopedic operation 67 (49.3%)
16 (11.8%) 9 (9.5%) 7 (17.1%) 0.207
Sickle cell disease 14 (10.3%) 8 (8.4%) 6 (14.6%) 0.274
Exogenous fat-soluble substances injection 3 (2.2%) 1 (1.1%) 2 (4.9%) 0.449
Renal angiomyolipoma 2 (1.5%) 2 (2.1%) 0 1.000 Lung transplantation 2 (1.5%) 1 (1.1%) 1 (2.5%) 1.000 Coronary artery bypass graft 1 (0.7%) 1 (1.1%) 0 1.000 Hepatocellular carcinoma 1 (0.7%) 0 1 (2.4%) 0.301 Multiple soft tissue injuries 1 (0.7%) 0 1 (2.4%) 0.301 Hypertriglyceridemia 1 (0.7%) 1 (1.1%) 0 1.000
Pre-existing conditions 53 (39.0%)
31 (32.6%) 22 (53.7%) 0.021
Hypertension 11 (8.1%) 7 (7.4%) 4 (9.8%) 0.900 Coronary artery disease 1 (0.7%) 1 (1.1%) 0 1.000 Arrhythmia 4 (2.9%) 0 4 (10.8%) 0.026 Cerebrovascular disease 3 (2.2%) 2 (2.1%) 1 (2.4%) 1.000 Diabetes 5 (3.7%) 4 (4.2%) 1 (2.4%) 0.994 COPD 3 (2.1%) 2 (2.1%) 1 (2.4%) 1.000 Asthma 3 (2.2%) 2 (2.1%) 1 (2.4%) 1.000 Connective tissue disease 4 (2.9%) 3 (3.2%) 1 (2.4%) 1.000 Tumor 4 (2.9%) 4 (4.2%) 0 0.315 Renal dysfunction 2 (1.5%) 1 (1.1%) 1 (2.4%) 1.000 Alzheimer’s disease 2 (1.5%) 0 2 (4.9%) 0.089 Drug abuse & drinking 3 (2.2%) 1 (1.1%) 2 (2.2%) 0.449
Diagnostic evaluations Chest CT
Ground-glass opacities 19/45 (42.2%)
11/31 (35.5%)
7/32 (21.9%)
2/44 (4.6%)
Table 1 (continued )
Non- survivor
pressure 29/34 (85.3%)
Glucocorticoid therapy 22 (16.2%)
Anticoagulant therapy 5 (3.7%) 4 (4.2%) 1 (2.4%) 0.994
Neurological therapy 13 (9.6%) 11 (11.6%) 2 (4.9%) 0.367
COPD indicates chronic obstructive pulmonary disease; orthopedic operation indicates total hip or knee arthroplasty, internal fixation and spinal instru- mentation. N, numbers; P value indicates the level of statistical significance between survivors and non-survivors; Q1-Q3, first to third quartile. Right ven- tricular abnormality is defined as right ventricular hypokinesis, and/or dilation, and/or dysfunction; Left ventricular abnormality is defined as left ventricular hypokinesis, and/or dysfunction; CT indicates computed tomography; Neuro- logical therapy is defined as sedation, and/or anti-convulsion, and/or inter- ventional thrombectomy; P values indicate differences between survivor and non-survivor patients. P < 0.05 was considered statistically significant.
Z. He et al.
114
Fig. 1. Clinical features and therapies in patients with fat embolism syndrome. A) Relative proportion of clinical manifestation observed in patients with FES. B) Frequencies of each clinical manifestation in over-all cases. C) Relative proportion of diagnostic findings that guided the final diagnosis of FES in over-all cases. D) Frequencies of auxiliary examination manifestations in over-all cases. E) Frequencies of respiratory supportive treatment, corticosteroid treatment, anticoagulant treatment and neurological treatment in over-all patients, survivors and non-survivors. *P = 0.036 for non-survivors vs survivors. F) 20 (21.05%) patients received corticosteroid treatment in survival group and 2 (4.88%) patients received corticosteroid in non-survival group. G) Constitute ratio for the use of each corticosteroid in patients of FES. H) Constitute ratio for the dose of corticosteroid used in patients of FES.
Z. He et al.
115
3.4. Diagnostic evaluation
Among 136 patients with fat embolism, 89 (65.4%) cases were diagnosed by clinical features combined with imaging findings, 29 (21.3%) cases were confirmed by autopsy using fat staining, 16 (11.8%) cases were diagnosed only by clinical features and the remaining 2 (1.5%) cases were confirmed by interventional thrombectomy (Fig. 1C). Chest CT-scan was a commonly reported diagnostic exam, performed in 45 (33.1%) patients. The most common finding on CT scan was ground- glass opacities in 19 (42.2%) patients, followed by normal findings in 10 (22.2%) cases, and patchy infiltration shadow in 6 (13.3%) patients (Fig. 1D). Chest X-ray was reported in 39 (27.9%) patients, with the most common manifestation being a patchy infiltration shadow, seen in 16 (42.1%) cases, followed by a decrease in transmittance seen in 9 (23.7%) (Fig. 1D). Among the 45 (33.1%) patients who underwent brain CT scan, 15 (33.3%) patients had low-density shadow while 21 (46.7%) had normal findings (Fig. 1D). Among 60 (44.1%) patients who under- went echocardiography, 11 (18.3%) had right ventricular hypokinesis and/or dilation and/or dysfunction, 9 (15%) had pulmonary hyperten- sion, 3 (5%) had left ventricular hypokinesis and/or dysfunction, 1 (1.7%) had a patent foramen ovale (Fig. 1D). Of 34 (25.0%) patients who had an arterial blood gas analysis, 29 (85.3%) had reduced oxygen partial pressure (Fig. 1D). Of 48 (35.3%) patients who had available blood tests, 33 (68.8%) had anemia and 28 (58.3%) had thrombocyto- penia (Fig. 1D). No significant difference was observed in the manifes- tation of these exams in survivors vs. non-survivors (Table 1).
3.5. Therapy
Most patients received respiratory supportive care, reported in 127 (93.4%) patients. Other therapies included corticosteroids, reported in 22 (16.2%) patients, while only 5 (3.7%) patients received anti- coagulation therapy and 13 (9.6%) patients received neurologic therapy (Fig. 1E, Table 1). When comparing the administered therapies between
survivors and non-survivors, more survivors received corticosteroid therapy compared with non-survivors, respectively 20 (21.1%) vs. 2 (4.9%) patients (p = 0.036), while there was no significant difference in the use of other treatments between two groups (Fig. 1E, and Table 1). We further compared mortality between patients with and without corticosteroids. Mortality in patients who received corticosteroids was significantly lower compared with that in patients who did not receive corticosteroids: 2/22 (9.1%) versus 39/114 (34.2%) patients, respec- tively (p = 0.036) (Fig. 1F). Of 22 patients who had been treated with corticosteroids, 11 received methylprednisolone, 3 prednisone, 2 dexa- methasone, 1 hydrocortisone, whereas in the other 5 patients the type of corticosteroid was not mentioned (Fig. 1G). Eight patients had been treated with high doses of corticosteroids (6 patients were treated with methylprednisolone at a dosage of 250 mg or higher, whereas in 2 pa- tients it has been mentioned that high dosage was used without reporting the definite dose). Eight patients had been treated with a lower dose of corticosteroid (less than 150 mg methylprednisolone or other equally efficient corticosteroid), while in the remaining 6 patients, data on the dosage of corticosteroids were not available (Fig. 1H).
3.6. Factors associated with in-hospital prognosis
Univariate and multivariate analyses revealed that age equal or above 65 years and non-orthopedic associated condition were signifi- cantly associated with increased mortality with an OR of 4.816 (95%CI 1.638–14.160, P = 0.004) and 4.785 (95%CI 1.019–22.474, P = 0.047) at multivariate analysis, while the use of corticosteroids was signifi- cantly associated with a reduced mortality with an OR of 0.143 (95%CI 0.029–0.711, P = 0.017) (Table 2).
4. Discussion
The major finding of this single-case metanalysis of published cases of FES is that in-hospital mortality is still unacceptably high, with a
Fig. 2. Images of femoral X-ray, computed tomographic pulmonary angiography and plain chest CT scan of a previously unpublished case with fat embolism syndrome (FES). The images were from a 31-year-old woman with FES. The patient complained of dyspnea and chest tightness for 2 days after a fall. A) X-ray of the femur demonstrating a fracture of the middle-left femur. Blood tests revealed anemia with increased d-dimer levels of 3.54 μg/ml. B) A chest computed tomography (CT) angiography ruled out the diagnosis of pulmonary artery embolism, as there was no obvious filling defect in the main pulmonary artery and bilateral pulmonary artery branches showed in mediastinal window (left); but the lung window showed multiple patches in both lungs, with blurred boundaries and partial fusion (right). The patient was initially treated as pneumonia and respiratory failure without clinical benefit. Further diagnostic evaluations were performed. C)…