Author + information
- Received September 28, 2013
- Revision received December 10, 2013
- Accepted January 21, 2014
- Published online April 29, 2014.
- Alfonso Muriel, MSc∗,
- David Jiménez, MD, PhD†∗ (, )
- Drahomir Aujesky, MD‡,
- Laurent Bertoletti, MD§,
- Herve Decousus, MD§,
- Silvy Laporte, MD, PhD§,
- Patrick Mismetti, MD, PhD§,
- Francisco J. Muñoz, MD‖,
- Roger Yusen, MD¶,
- Manuel Monreal, MD, PhD#,
- RIETE Investigators
- ∗Biostatistics Unit, Ramón y Cajal Hospital and Instituto Ramón y Cajal de Investigación Sanitaria IRYCIS, CIBERESP, Madrid, Spain
- †Respiratory Department, Ramón y Cajal Hospital and Instituto Ramón y Cajal de Investigación Sanitaria IRYCIS, Madrid, Spain
- ‡Division of General Internal Medicine, Bern University Hospital, Bern, Switzerland
- §Thrombosis Research Group, Université de Saint-Etienne, Jean Monnet, Inserm, Service de Médecine Interne et Thérapeutique, Hôpital Nord, Saint-Etienne, France
- ‖Department of Internal Medicine, Hospital de Mollet, Barcelona, Spain
- ¶Divisions of Pulmonary and Critical Care Medicine and General Medical Sciences, Washington University School of Medicine, St. Louis, Missouri
- #Department of Internal Medicine, Hospital Universitari Germans Trias I Pujol, Badalona, Barcelona, Spain
- ↵∗Reprint requests and correspondence:
Dr. David Jiménez, Respiratory Department and Medicine Department, Ramón y Cajal Hospital, IRYCIS and Alcalá de Henares University, 28034 Madrid, Spain.
Objectives The purpose of this study was to investigate the survival effects of inferior vena cava filters in patients with venous thromboembolism (VTE) who had a significant bleeding risk.
Background The effectiveness of inferior vena cava filter use among patients with acute symptomatic VTE and known significant bleeding risk remains unclear.
Methods In this prospective cohort study of patients with acute VTE identified from the RIETE (Computerized Registry of Patients With Venous Thromboembolism), we assessed the association between inferior vena cava filter insertion for known significant bleeding risk and the outcomes of all-cause mortality, pulmonary embolism (PE)-related mortality, and VTE rates through 30 days after the initiation of VTE treatment. Propensity score matching was used to adjust for the likelihood of receiving a filter.
Results Of the 40,142 eligible patients who had acute symptomatic VTE, 371 underwent filter placement because of known significant bleeding risk. A total of 344 patients treated with a filter were matched with 344 patients treated without a filter. Propensity score–matched pairs showed a nonsignificant trend toward lower risk of all-cause death for filter insertion compared with no insertion (6.6% vs. 10.2%; p = 0.12). The risk-adjusted PE-related mortality rate was lower for filter insertion than no insertion (1.7% vs. 4.9%; p = 0.03). Risk-adjusted recurrent VTE rates were higher for filter insertion than for no insertion (6.1% vs. 0.6%; p < 0.001).
Conclusions In patients presenting with VTE and with a significant bleeding risk, inferior vena cava filter insertion compared with anticoagulant therapy was associated with a lower risk of PE-related death and a higher risk of recurrent VTE. However, study design limitations do not imply a causal relationship between filter insertion and outcome.
Despite the advances in the diagnosis and management of venous thromboembolism (VTE), deep vein thrombosis (DVT) and pulmonary embolism (PE) remain major causes of morbidity and mortality (1). Conventional treatment for VTE consists of the use of parenteral agents (i.e., unfractionated heparin [UFH], low-molecular-weight heparin [LMWH], fondaparinux) as a “bridge” for oral anticoagulation therapy (2). Guidelines do not recommend insertion of a filter in the inferior vena cava (IVC) as the primary treatment of VTE. A large population-based retrospective analysis, which assessed for recurrent VTE in patients treated with an IVC filter for acute VTE, found that the use of a filter was associated with a higher incidence of rehospitalization for venous thrombosis among patients who initially manifested PE (3). Stein et al. (4) showed that the all-cause in-hospital case fatality rate was lower among patients with unstable PE who received thrombolytic therapy and had a vena cava filter. In the only clinical trial that evaluated the efficacy of vena cava filters (in addition to standard anticoagulant therapy), this treatment reduced the risk of PE but increased that of DVT and had no effect on survival (5). An 8-year follow-up of the patients enrolled in the PREPIC (Prevention of Recurrent Pulmonary Embolism by Vena Cava Interruption) trial showed similar results (6).
In the absence of randomized clinical trials that demonstrate a mortality benefit of IVC filter treatment of VTE, the American College of Chest Physicians guidelines mainly limit their recommendation for IVC filter insertion to patients with acute symptomatic VTE and a contraindication to anticoagulation (grade 1B) (2). Unfortunately, studies have not clearly determined which patients with VTE would benefit from vena cava filter therapy.
Given the lack of data supporting a survival benefit of IVC filter therapy in patients with acute VTE, we conducted the present study using data collected for an international multicenter (Online Appendix) (7,8). The study assessed the association between the insertion of an IVC filter and mortality and other outcomes during the first month after treatment for acute symptomatic VTE in patients who had known significant bleeding risk.
This retrospective study used prospectively collected data from patients enrolled in the RIETE (Computerized Registry of Patients With Venous Thromboembolism) (Online Appendix) (7,8). All patients provided written or oral informed consent for participation in the registry in accordance with local ethics committee requirements.
Study cohort and definition of treatment groups
At each participating site, RIETE investigators aimed to enroll consecutive patients who had acute symptomatic or asymptomatic VTE confirmed by using objective testing that consisted of: high-probability ventilation/perfusion scintigraphy (9); positive contrast-enhanced PE protocol; helical chest computed tomography (CT) (single or multidetector CT) for PE (10); or lower limb venous compression ultrasonography positive for proximal DVT (11). This study excluded those who had asymptomatic VTE. Of the patients treated with an IVC filter, this study only included those patients who had an IVC filter inserted during the first 30 days after VTE diagnosis because of known significant bleeding risk (i.e., absolute or relative contraindication to anticoagulation therapy) as determined by the local investigator (i.e., not independently adjudicated).
Treated patients were defined as those who received an IVC filter (with or without concomitant anticoagulation) because of known significant bleeding risk. Control patients were defined as those with distributions of observed baselines covariates (i.e., similar baseline risk of bleeding) comparable to treated patients (see the Statistical analysis section) who did not receive a filter but underwent anticoagulant therapy.
Patients enrolled in the RIETE registry had data collected from around the time of VTE diagnosis that included but was not limited to: age; sex; weight; presence of coexisting conditions such as chronic heart or lung disease; recent (<30 days before VTE) major bleeding; presence of risk factors for PE, including active cancer (defined as newly-diagnosed cancer or cancer being treated [i.e., surgery, chemotherapy, radiotherapy, hormonal or support therapy]); recent immobility (defined as nonsurgical patients assigned to bed rest with bathroom privileges for ≥4 days in the 2 months before VTE diagnosis); surgery (defined as those who had undergone major surgery in the 2 months before VTE); clinical signs and symptoms on admission, including heart rate and systolic blood pressure; and laboratory results at hospital admission that included hemoglobin (Hb), platelet count, and serum creatinine.
This study used all-cause mortality through 30 days after initiation of anticoagulant treatment or filter insertion as the primary endpoint, and 30-day PE-related mortality, recurrent VTE, and major bleeding as secondary endpoints. The RIETE investigators assessed mortality presence, cause, and date by using medical record review and proxy interviews when necessary. Clinicians at RIETE-enrolling sites managed patients with suspected recurrences according to their local practice. Typically, the RIETE investigators defined recurrent DVT as a new noncompressible vein segment or an increase of the vein diameter by at least 4 mm compared with the last available measurement on venous ultrasonography (12); recurrent PE as a new ventilation/perfusion mismatch on lung scan or a new intraluminal filling defect on spiral CT of the chest (10); and major bleeding episodes as those that required a transfusion of at least 2 U of blood, were retroperitoneal, spinal, or intracranial, or were fatal (13).
Treatment and follow-up
Clinicians at RIETE-enrolling sites managed patients according to their local practice (i.e., there was no standardization of treatment). In the RIETE registry, most patients received initial anticoagulation with intravenous UFH or subcutaneous LMWH or fondaparinux, and overlap and long-term therapy with an oral vitamin K antagonist. Clinicians administered thrombolytic treatment and/or inotropic support as deemed appropriate. In general, clinicians used thrombolytic treatment of acute PE in patients with cardiogenic shock (e.g., persistent systolic arterial pressure <90 mm Hg in the setting of clinical signs of organ hypoperfusion [clouded sensorium, oliguria, cold and clammy skin, or lactic acidosis]). The RIETE registry recorded information related to patient outcomes through 3 months after the diagnosis of acute VTE, and this study analyzed outcomes through 30 days after initiation of VTE treatment (i.e., anticoagulation or filter insertion).
Chi-square or Fisher exact tests were used to compare categorical data between groups. The Kolmogorov-Smirnov test was used to assess continuous data for a normal distribution. We used 2-tailed unpaired Student t tests to compare normally distributed continuous data between 2 groups, and the Mann-Whitney U test was used for non-normally distributed continuous data comparisons. Multivariate adjustments were made for age, cancer, recent or active bleeding, immobilization, heart rate, systolic blood pressure, creatinine levels, platelet count, and Hb levels via logistic regression to see if insertion of a filter was an independent significant predictor of all-cause mortality in the entire sample (N = 40,142).
Because clinicians did not randomly allocate filter therapy, the patients who received an IVC filter likely systematically differed from patients who did not receive one with respect to baseline characteristics, clinical course, clinical examination and test findings, and comorbid conditions. A propensity score adjustment was used to compare treatment effects for patients with similar predicted probabilities of receiving a filter (14). Logistic regression was used to estimate propensity scores, and we modeled the log odds of the probability that a patient received a filter by using baseline demographic and clinical variables (see the Baseline variables section for definitions) that were previously shown to be associated with mortality or treatment selection. These variables included: patient age at time of diagnosis of VTE; presence or absence of recent (<30 days before VTE) major bleeding; presence or absence of active bleeding at the time of diagnosis of VTE; presence or absence of risk factors for VTE that included active cancer or recent immobility; clinical signs and symptoms on admission, including heart rate and systolic blood pressure; and anemia (defined as Hb level <13 g/dl in men and <12 g/dl in women) (15), thrombocytopenia (platelet count <100 × 109/l), and abnormal serum creatinine levels (>2 mg/dl).
After generation of the propensity scores, we sought to estimate the reduction in 30-day overall mortality attributable to the insertion of a filter by using a greedy matched-paired analysis that has a 1:1 matching algorithm and does not allow for replacements. We randomly selected a patient in the treatment group and then matched that patient with the nearest patient in the control group within a fixed caliper width of 0.05 (16). To assess the success of the matching procedure, we measured standardized differences (measured in percentage points) in observed confounders between the matched groups (17). The filter effect was estimated by using generalized estimating equation methods to incorporate the matched-pairs design, and we adjusted for those covariates that remained unbalanced after matching (18). The data were analyzed according to 3 different propensity score models: 1 for any VTE, 1 for symptomatic DVT (without symptomatic PE), and 1 for symptomatic PE. The area under the curve (AUC) was calculated for each receiver-operating characteristic curve to quantify propensity score model accuracy and its ability to correctly discriminate or identify those patients who received and those who did not receive an IVC filter.
Several sensitivity analyses were performed. First, we examined the effect of filters inserted during the first 7 days after the diagnosis of VTE. Because most fatal PEs occur within the first few days after VTE diagnosis (8), we expected that the effectiveness of filters would be greatest soon after the VTE diagnosis. Second, we evaluated various caliper widths iteratively until between-group standardized differences were minimized. Finally, we repeated all analyses for the secondary endpoints.
We used psmatch2 for the propensity score analyses and Stata version 11.2 for Windows (StataCorp, College Station, Texas) for all other analyses.
A total of 40,975 patients were identified with objectively-confirmed VTE enrolled in the RIETE registry during the study period. After the exclusion of 230 (0.6%) patients with asymptomatic VTE and 603 (1.5%) patients who received a filter for reasons other than a relative or absolute contraindication to anticoagulation, the study cohort consisted of 40,142 patients with confirmed DVT (51%), PE (32%), or both (17%) (Fig. 1).
Of the 371 patients who received an IVC filter, 152 (41%) had DVT (without symptomatic PE) and 219 had PE (with or without symptomatic DVT). Patients treated with filters and those treated without filters differed significantly in pre-existing medical conditions and relevant clinical, physiological, and laboratory parameters (Tables 1 and 2). Patients who received a filter had more comorbid diseases (cancer, recent surgery, immobility, congestive heart failure, chronic lung disease, and recent bleeding), signs of clinical severity (tachycardia and hypotension), and laboratory abnormalities (renal failure, anemia, and thrombocytopenia) compared with those who did not receive a filter.
Overall, 1,823 (4.5%) of 40,142 patients died (all-cause mortality) through 30 days after the diagnosis of VTE. Three percent of the patients with DVT (602 of 20,503 patients) died during the 1-month follow-up period, compared with 6.2% of the patients with PE (1,221 of 19,639 patients). Of the patients who received an IVC filter, 7.3% (27 of 371 patients) died during the 30-day follow-up period. Of those who did not receive a filter, 4.5% (1,796 of 39,771 patients) died (absolute difference [AD]: 2.8% [95% confidence interval (CI) of the AD: 0.5% to 5.9%]; p = 0.02) during follow-up. Of the DVT patients (without symptomatic PE) who received an IVC filter, 5.9% died during the follow-up period, compared with 2.9% of those who did not receive a filter (AD: 3.0% [95% CI of the AD: 0.2% to 8.0%]; p = 0.05) during follow-up. Of the PE patients, with or without DVT, who received an IVC filter, 8.2% died, whereas 6% of those who did not receive a filter died (AD: 2.0% [95% CI of the AD: 0% to 6.4%]; p = 0.28) during follow-up.
The matching of patients presenting with any acute VTE yielded 344 patients treated with filters and 344 patients treated without filters. This model showed good to excellent discrimination, with an AUC of 0.80. The standardized differences of <10% for all matched variables supported the assumption of balance between treatment groups in observed confounders (Table 3). We successfully matched 134 patients with acute symptomatic DVT who received a filter with 134 patients who did not on the basis of propensity score (AUC: 0.87). The matching process eliminated some significant differences that existed between groups regarding pre-existing medical conditions or relevant clinical, physiological, and laboratory parameters (Table 4). Propensity analyses of the subgroup of patients with PE used 210 matched pairs (210 patients from each group) (AUC: 0.74). The matched sample showed good balance for each variable.
Among those patients with DVT who did not have a filter, 128 received anticoagulant therapy and 6 did not. Of the 128 anticoagulated patients with DVT, 121 patients (95%) received therapy with LMWH, 4 (3%) received UFH, and 3 (2%) received fondaparinux. Among those patients with DVT who had a filter inserted, 80 received anticoagulants and 54 were not anticoagulated. A total of 67 patients (84%) received therapy with LMWH, 12 (15%) received UFH, and 1 (1%) received fondaparinux. Figure 2 shows LMWH dosing regimens among matched patients who were treated with or without filters.
Among those patients with PE who did not have a filter, 200 received anticoagulant therapy and 10 were not anticoagulated. Of the 200 anticoagulated patients with PE, 174 patients (87%) received therapy with LMWH, 23 (12%) received UFH, and 3 (1%) received fondaparinux. Among those patients with PE who had a filter inserted, 118 received anticoagulants and 92 were not anticoagulated. A total of 84 patients (71%) received therapy with LMWH, 31 (26%) received UFH, and 3 (3%) received fondaparinux. Figure 2 shows the LMWH dosing regimens among matched patients who were treated with or without filters.
Filter insertion was associated with a nonstatistically significant lower mortality rate than nonfilter treatment in the matched cohort of patients with any VTE (6.6% vs. 10.2%; risk difference: –3.6% [95% CI: –7.7% to 0.7%]; p = 0.12). A similar reduction in 30-day all-cause mortality was evident for both subtypes of VTE (DVT and PE) (Table 5). Analysis of propensity score–matched pairs (n = 344 pairs) showed a statistically significant decreased risk of PE-related mortality for filter insertion compared with no insertion (1.7% vs. 4.9%; risk difference: –3.2% [95% CI: –6.2% to –0.5%]; p = 0.03).
After propensity score matching, there was no significant difference in the rate of major bleeding at 30 days between patients receiving filters and those not receiving filters (3.8% vs. 5.2%; risk difference: –1.4% [95% CI: –1.7% to 4.5%]; p = 0.35) (Table 5). This difference was also nonstatistically significant in the matched cohort of patients who initially manifested PE (3.8% vs. 6.7%; risk difference: –2.9% [95% CI: –1.4% to 7.2%]; p = 0.18). In the subgroup of patients with DVT, the number of events did not allow for comparisons.
The rates of recurrent VTE at 30 days in the matched cohort of patients with any VTE were significantly higher among patients receiving filters than among those not receiving filters (6.1% vs. 0.6%; risk difference: 5.5% [95% CI: 2.8% to 8.2%]; p < 0.001). Of the 21 recurrent events that occurred in patients who received a filter, 15 (71%) were DVT and 6 (29%) were PEs. Two patients who did not receive a filter experienced a PE recurrence during follow-up. In the matched cohort of patients with PE, the recurrent VTE rate was significantly higher in those patients who received an IVC filter versus those who did not (8.1% vs. 1.0%; risk difference: 7.1% [95% CI: 3.2% to 11.0%]; p < 0.001). In the subgroup of matched patients with DVT, the number of events did not allow for comparisons.
When we evaluated filters inserted during the first 7 days after the diagnosis of VTE, results mimicked the findings of the primary analysis: mortality rates in the group receiving filters tended to be lower than those in the group not receiving filters, whereas VTE recurrences were significantly higher. Similarly, we found consistent results when we evaluated various caliper widths (0.1 and 0.2 SD).
Uncertainty surrounding the efficacy of IVC filters exists. The results of our study show that, in patients with significant bleeding risk, filter insertion was associated with a lower risk of 30-day PE-related mortality compared with anticoagulant therapy, whereas it was associated with a higher risk of recurrent VTE. IVC filter therapy had a similar effect on each of the study endpoints in the subgroups of patients who presented solely with DVT or with PE (with or without DVT). However, study design limitations do not imply a causal relationship between filter insertion and outcome.
Despite the relatively frequent use of IVC filters, few prospective, randomized controlled trials have assessed their efficacy and safety in patients with acute symptomatic VTE (19). Decousus et al. (%) evaluated the efficacy of permanent vena cava filters in the PREPIC study, in which 400 patients with proximal DVT were randomly selected to receive a permanent vena cava filter (6). The study also randomized patients to receive 1 of 2 classes of anticoagulants. The study did not detect a difference in the death rate at any time during follow-up when comparing those with and without vena filters. Permanent filters reduced the risk of PE but increased the risk of DVT (5). Although this study provided helpful information, it did not address the value of IVC filters versus anticoagulant therapy in those situations in which filters are commonly advocated (i.e., absolute contraindication to anticoagulation).
Many studies have demonstrated the relative efficacy and safety of standard anticoagulant therapy in hemodynamically-stable patients with PE. For patients with VTE and known significant bleeding risk, this study suggested that use of an IVC filter, compared with anticoagulant treatment, was associated with fewer deaths attributable to PE. Patients treated with an IVC filter did not receive anticoagulation therapy or received lower doses of anticoagulants than those not treated with a filter. Thus, filter insertion may have prevented some fatal PE recurrences. However, the IVC filter group had a higher rate of recurrent symptomatic VTE compared with the nonfilter group. Lack of anticoagulant treatment or thrombosis at the filter site may have caused the higher recurrent DVT rate. Contrary to what might have been expected, IVC filters increased the rate of symptomatic PE recurrence. Because patients treated with an IVC filter had a lower PE-related mortality than those not treated with a filter, we speculate that filters did not allow large PE to recur.
The risk of bleeding and other variables in an individual patient should lead to tailoring of the intensity and type of treatment of symptomatic VTE (20). In this study, clinicians withheld or used lower doses of anticoagulants in patients treated with an IVC filter, and this approach had a nonstatistically significant lower 30-day major bleeding rate compared with standard anticoagulation. These findings occurred even though the filter-treated subgroup had more risk factors for bleeding than the anticoagulated cohort. Predicting which patients will have major bleeding during anticoagulant therapy after VTE remains difficult. Researchers have developed decision tools to better define the risk of anticoagulant-associated bleeding (21–23). To determine whether such models can help clinicians decide which treatment options to choose will require further prospective studies in various at-risk groups.
Using data from the Nationwide Inpatient Sample, Stein et al. (4) reported decreased case fatality rates with vena cava filters in patients in unstable condition, regardless of whether they received thrombolytic therapy, and in stable patients who received thrombolytic therapy. However, to the best of our knowledge, no other population-based study has reported on the effectiveness of IVC filter use as an acute management strategy for patients with acute VTE and known significant bleeding risk. According to a scientific statement from the American Heart Association (24), adult patients with any confirmed acute PE (or proximal DVT) with contraindications to anticoagulation or with active bleeding complications should receive an IVC filter. Results of the present investigation support this recommendation. An ongoing multicenter randomized trial (PREPIC II [Prevention of Recurrent Pulmonary Embolism by Vena Cava Interruption [PREPIC2]]; NCT00457158) aims to determine the efficacy and safety of insertion of an IVC filter in the prevention of PE recurrence (25).
Selection bias could have skewed the study sample. However, the broad range of patients with acute symptomatic VTE from multiple medical centers, countries, and treatment settings enrolled in the RIETE registry decreased the likelihood of the inclusion of a skewed population in this study. The population-based sample we used described the effects of filters in “real-world” clinical care and enhanced the generalizability of the findings. Confounding may have affected the results of this observational study. We used propensity score matching to make the patient groups comparable according to the measured confounders (i.e., baseline risk of bleeding), and we successfully eliminated the observed differences. However, residual confounding may still have occurred. The similar results from many of the sensitivity and secondary analyses provided evidence of the robustness of the findings and further strengthened the soundness of the conclusions. Despite the large number of patients assessed for this study from the RIETE registry, the relatively small sample size of the propensity-matched cohorts lowered the statistical power of the study and therefore increased the chance that the study would not detect a statistically significant difference in outcomes between the treatment groups (i.e., type II error). Finally, concomitant anticoagulation might have accounted for some of the differences between the groups' outcomes. However, because clinicians typically use lower doses of anticoagulants in patients with a high risk of bleeding, the study findings suggest that the design might have caused a slight bias against filter insertion.
In patients with acute symptomatic VTE and known significant bleeding risk, IVC filter therapy may reduce the risk of PE-related mortality compared with anticoagulant therapy. However, lack of anticoagulation or thrombosis associated with IVC filters could increase the risk of recurrent VTE. Randomized controlled trials of retrievable filters might help to identify subgroups of patients at high risk of death who might have a favorable risk to benefit ratio for treatment with filter insertion.
The authors thank the Registry Coordinating Center, S & H Medical Science Service, for their quality control, logistic, and administrative support.
The RIETE registry was supported by an unrestricted educational grant from Sanofi Spain and by Bayer Pharma AG. (Bayer Pharma AG's support was limited to the part of the RIETE registry outside of Spain, which accounts for approximately 18% to 19% of the total patients.) Dr. Yusen has received grants from Bayer HealthCare Pharmaceuticals, Inc., Portola, Inc., Pfizer, Inc., and Bristol-Myers Squibb; has received consulting fees from Bayer HealthCare, Bristol-Myers Squibb, and GlaxoSmithKline; has served as a legal consultant for Ortho Pharmaceuticals, Inc., Organon, Inc., Pfizer, Portola, and Sanofi-Aventis; and was a member of the Data Safety Monitoring Board for a National Institutes of Health, National Heart, Lung, and Blood Institute–funded trial. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Drs. Muriel and Jiménez contributed equally to this work.
- Abbreviations and Acronyms
- absolute difference
- area under the curve
- confidence interval
- computed tomography
- deep vein thrombosis
- inferior vena cava
- low-molecular-weight heparin
- pulmonary embolism
- unfractionated heparin
- venous thromboembolism
- Received September 28, 2013.
- Revision received December 10, 2013.
- Accepted January 21, 2014.
- American College of Cardiology Foundation
- Kearon C.,
- Akl E.A.,
- Comerota A.J.,
- et al.,
- American College of Chest Physicians
- Decousus H.,
- Leizorovicz A.,
- Parent F.,
- et al.
- PREPIC Study Group
- Laporte S.,
- Mismetti P.,
- Décousus H.,
- et al.,
- RIETE Investigators
- Prandoni P.,
- Cogo A.,
- Bernardi E.,
- et al.
- World Health Organization
- Austin P.C.
- Austin P.C.,
- Mamdani M.M.
- Jaff M.R.,
- McMurtry M.S.,
- Archer S.L.,
- et al.,
- American Heart Association Council on Cardiopulmonary, Critical Care, Perioperative and Resuscitation; American Heart Association Council on Peripheral Vascular Disease; American Heart Association Council on Arteriosclerosis, Thrombosis and Vascular Biology
- Mismetti P.