Intensive Care Med (2013) 39:872880 DOI 10.

1007/s00134-013-2814-2

ORIGINAL

Philippe Vignon Pierre-Franc ois Dequin Anne Renault Armelle Mathonnet Nicolas Paleiron Audrey Imbert Delphine Chatellier rie Gissot Vale ritier Gwenaelle Lhe Victor Aboyans Gwenael Prat Denis Garot Thierry Boulain Jean-Luc Diehl Luc Bressollette lien Delluc Aure Karine Lacut The Clinical Research in Intensive Care and Sepsis Group (CRICS Group)
Received: 20 July 2012 Accepted: 6 January 2013 Published online: 31 January 2013 Springer-Verlag Berlin Heidelberg and ESICM 2013

Intermittent pneumatic compression to prevent venous thromboembolism in patients with high risk of bleeding hospitalized in intensive care units: the CIREA1 randomized trial

N. Paleiron Service des maladies respiratoires, HIA Clermont-Tonnerre, Brest, France A. Imbert J.-L. Diehl animation Me dicale, Re HEGP AP-HP, Paris, France D. Chatellier animation Me dicale, Service de Re CHU de Poitiers, Poitiers, France

K. Lacut INSERM, CIC 0502, Brest, France K. Lacut ()) pital de la Cavale Blanche, CIC 0502, Ho Bd Tanguy Prigent, F-29609 Brest Cedex, France e-mail: karine.lacut@chu-brest.fr Tel.: ?33-29-8145007 Fax: ?33-29-8145008

ritier P. Vignon G. Lhe animation Polyvalente, CHU Dupuytren, Re 87042 Limoges, France ritier P. Vignon G. Lhe INSERM, CIC-P 0801, CHU Dupuytren, 87042 Limoges, France ritier P. Vignon G. Lhe de Limoges, Universite 87000 Limoges, France P.-F. Dequin D. Garot Franc Universite ois Rabelais et service animation Me dicale, de Re pital Bretonneau, CHU de ToursHo Tours, France A. Renault G. Prat animation Me dicale, Service de Re CHU de la Cavale Blanche, Brest, France A. Mathonnet T. Boulain animation, CHRHo pital de la Source, Re ans, France Orle

Abstract Purpose: Venous thromboembolism (VTE) is a frequent and serious problem in intensive care units (ICU). AnticoagV. Aboyans ulant treatments have demonstrated Service de Cardiologie, CHU de Limoges, their efcacy in preventing VTE. Limoges, France However, when the bleeding risk is high, they are contraindicated, and J.-L. Diehl Paris Descartes, mechanical devices are recomINSERM U765, Universite , France Sorbonne Paris Cite mended. To date, mechanical prophylaxis has not been rigorously L. Bressollette A. Delluc K. Lacut evaluated in any trials in ICU de Bretagne EA3878, Universite patients. Methods: In this multiOccidentale, 29200 Brest, France center, open-label, randomized trial with blinded evaluation of endpoints, L. Bressollette de Me decine Vasculaire, Unite we randomly assigned 407 patients CHU Brest, 29609 Brest, France with a high risk of bleeding to receive intermittent pneumatic compression A. Delluc (IPC) associated with graduated Departement de Medecine Interne compression stockings (GCS) or GCS et Pneumologie, CHU Brest, alone for 6 days during their ICU 29609 Brest, France
V. Gissot animation Polyvalente, Service de Re me, Saint-Michel, France CH DAngoule

873

stay. The primary endpoint was the occurrence of a VTE between days 1 and 6, including nonfatal symptomatic documented VTE, or death due to a pulmonary embolism, or asymptomatic deep vein thrombosis detected by ultrasonography systematically performed on day 6. Results: The primary outcome was assessed in 363 patients (89.2 %). By day 6, the incidence of the primary outcome was 5.6 % (10 of 179

patients) in the IPC ? GCS group and 9.2 % (17 of 184 patients) in the GCS group (relative risk 0.60; 95 % condence interval 0.281.28; p = 0.19). Tolerance of IPC was poor in only 12 patients (6.0 %). No intergroup difference in mortality rate was observed. Conclusions: With the limitation of a low statistical power, our results do not support the superiority of the combination of IPC ? GCS compared to GCS alone

to prevent VTE in ICU patients at high risk of bleeding. Keywords Intermittent pneumatic compression devices Elastic stockings Venous thromboembolism Venous thrombosis Intensive care units

trial to determine the efcacy and safety of IPC associated with GCS versus GCS alone for the prevention of VTE in Venous thromboembolism (VTE), that occurs as either ICU patients with a high risk of bleeding. deep vein thrombosis (DVT) or pulmonary embolism (PE), is a major cause of morbidity and mortality in patients admitted to intensive care units (ICU) [1]. CritiMaterials and methods cally ill patients have often multiple risk factors for VTE [2, 3]. Most of these risk factors are related to age or co Trial design morbidities (e.g., cancer, obesity), or are acquired during the ICU stay (e.g., central venous catheters, mechanical CIREA 1 (Compression pneumatique Intermittente en ventilation, or vasopressor use) [24]. The reported inci- REAnimation) was a multicenter, open-label, randomdence rates of VTE in hospitalized patients range between ized, outcome-blinded trial conducted at nine ICUs in 20 % and 80 % depending to the studied population, the France (see Appendix). The trial protocol was approved type of thromboprophylaxis and the screening test used by the local institutional review board. The sponsor [5]. In the ICU setting, PE may account for acute episodes played no role in the analysis or drafting of the of hemodynamic instability or hypoxia and may contrib- manuscript. ute to failure of weaning from mechanical ventilation [6, 7]. Since VTE is frequently silent and clinically unsuspected in ICU patients [810], PE remains one of Study population the most common unexpected autopsy ndings in the critically ill [11], and is found at autopsy in 7 % to 27 % Patients aged 18 years or older who were at high risk of of ICU fatalities [12, 13]. bleeding on ICU admission were eligible for the trial. A Surprisingly, little information is yet available on VTE high risk of bleeding was dened as symptomatic bleedprophylaxis in the ICU. Randomized trials including ICU ing or the presence of organic lesions likely to bleed, patients have indicated a benet of unfractionated or low hemophilic diseases, hemostatic abnormalities (platelet molecular weight heparin over placebo, without signi- count \50,000/mm3, aPTT ratio [2, prothrombin time cant risk of hemorrhage [1416]. In ICU patients who are \40 %), or the presence of severe anemia (hemoglobin at high risk of bleeding, anticoagulant treatments are \7 g/dl) due to bleeding or unexplained. Prolonged contraindicated, and current guidelines recommend the clotting times can be the result of the use of anticoaguoptimal use of mechanical thromboprophylaxis with lants. These parameters alone when not associated with graduated compression stockings (GCS) and/or intermit- bleeding or a high risk of bleeding could not be considtent pneumatic compression (IPC) at least until the ered as inclusion criteria. Other requirements for bleeding risk decreases [17]. IPC has been shown to be inclusion were a willingness to participate for the duration effective in preventing VTE in neurosurgical patients of the trial, and the provision of written informed consent undergoing craniotomy [18]. A randomized trial con- given by the patients next of kin. ducted in patients with intracerebral hemorrhage has Exclusion criteria were patient refusal, the absence of shown that IPC is more efcient in preventing asymp- a high risk of bleeding as previously dened, the presence tomatic DVT than GCS alone [19]. However, the efcacy of a documented VTE at screening or a recent DVT and safety of mechanical prophylaxis of VTE have not yet (\3 months), ICU stay of more than 36 h or likely to be been specically evaluated in ICU patients. Accordingly, \72 h, a life-support limitation, a contraindication for we carried out an open-label randomized parallel-group mechanical prophylaxis (i.e., severe lower limb

Introduction

874

arteriopathy, any arterial graft of the legs, a wound in the criterion: (1) symptomatic DVT of a lower limb (objeclower limb related to either vascular disease or trauma), tively conrmed by CUS) or symptomatic, nonfatal and and the presence of a mechanical prosthetic heart valve. objectively conrmed PE, (2) death between day 1 and day 6 related to a PE, and (3) asymptomatic DVT (either distal or proximal) of a lower limb detected by CUS systematically performed on day 6. All reported sympStudy procedures tomatic VTE and all deaths that occurred between day 1 Patients were randomly assigned at ratio of 1:1 to receive and day 6 were judged on the basis of standardized crieither IPC associated with GCS or GCS alone. The ran- teria by an independent adjudication committee unaware domization was centralized with internet access (Clininfo, of the randomized treatment assignments. Only conrmed Lyon, France), and stratied by center and age group VTE and deaths considered as denitely related to PE were included in the analysis of the primary endpoint. The (1850 years, 5170 years, and [70 years). GCS consisted of thigh-length GCS (T.E.D.TM anti- CUS performed on day 6 by trained ultrasonographers embolism stockings; Covidien, Manseld, MA) and IPC was standardized [20] and was done blindly, i.e., the was achieved with the SCD EXPRESSTM compression patients had bare legs, and mechanical devices were system with adapted tubing sets and thigh sleeves (Co- removed and hidden before the procedure. All examinavidien). Nurses were trained in the use of mechanical tions were videotaped and sent to the coordination center devices to apply optimal compression (proper sizing of for validation by an expert ultrasonographer who was both GCS and sleeves for IPC, and their proper applica- unaware of the treatment group. DVT of the upper limb tion). GCS and IPC were applied to both legs as soon as was not considered as a primary endpoint. Secondary possible after randomization and maintained continuously endpoints were the occurrence of a symptomatic VTE until compression ultrasonography (CUS) was performed between day 6 and day 90, and death from any cause up to on day 6. After that day, the decision to maintain VTE day 30 or day 90. Adverse events were recorded during the use of prophylaxis and its modality were left at the discretion of the investigators. Anticoagulation was not permitted mechanical devices and the tolerability of GCS and IPC during the rst 6 days of the study. The use of GCS and was assessed using a specic chart. IPC was recorded to monitor compliance and tolerance. Compliance was considered poor if the mechanical devices were used less than 80 % of the time. The GCS Statistical analysis and IPC removal date and reasons were recorded when applicable. The use of procoagulant treatments (pro- Primary analysis was performed on an intention-to-treat thrombin complex concentrate, fresh frozen plasma, basis. Randomized patients who had a documented clinplatelet transfusion, vitamin K, coagulation factors, prot- ical endpoint or a CUS examination by day 6 were amine sulfate) between admission and day 6 was included in the efcacy endpoint analysis. The incidence of the composite primary endpoint on day 6 was comrecorded. Prior to randomization, CUS was recommended in pared between groups using an exact two-sided Fisher patients transferred to the ICU from another hospital or test. Relative risk and absolute risk reduction were comanother unit after hospitalization of at least 48 h. This puted with their 95 % condence intervals. Assuming an incidence of VTE of 15 % in the GCS baseline CUS was performed to exclude prevalent DVT (exclusion criterion). We assumed that patients admitted group [6, 19], we calculated that 356 patients were directly to the ICU for an acute event had no prevalent required for the study to detect a 60 % reduction in the DVT but those who were already hospitalized before incidence of VTE with the combination IPC ? GCS, with admission to the ICU could have developed DVT during a power of 80 % and a two-sided alpha level of 5 %. Because approximately 20 % of patients were expected to the rst days of hospitalization. All included patients were visited on day 6, and sub- die during the rst days of their ICU stay, we decided to sequently followed up on day 30 and day 90 (visit or enroll 400 patients. phone call). Patients or their relatives were questioned about VTE events after ICU discharge. If a patient died, the date and the most probable cause of death were Results recorded. Between 21 November 2007 and 20 December 2010 a total of 407 patients underwent randomization in the CIREA1 study, 205 in the IPC ? GCS group and 202 in the The primary endpoint with respect to efcacy was the GCS group (Fig. 1). The systematic CUS was performed incidence of VTE on day 6 assessed by a composite 5.8 1.6 days after ICU admission in the IPC ? GCS Endpoints

875

group and after 6.0 1.4 days in the GCS group. Three patients in the GCS group and one in the IPC ? GCS group received at least one dose of curative anticoagulant treatment before the CUS evaluation for symptomatic DVT of an upper limb (n = 2), atrial brillation (n = 1) and unspecied reason (n = 1). Six patients in the GCS group and one in the IPC ? GCS group received at least one dose of preventive anticoagulation treatment before the CUS evaluation. Baseline characteristics were well balanced between the two groups (Table 1). About half of the patients had a spontaneous or traumatic intracranial hemorrhage. In 29 patients (7.1 %), anticoagulant treatment was administered at a therapeutic dose before ICU admission (atrial brillation in 16, previous VTE [3 months in 7, thrombolysis for myocardial infarction in 2, antiphospholipid syndrome in 1, coronary disease in 1 and unknown reason

in 2) and 29 patients (7.1 %) received thromboprophylaxis, mainly mechanical thromboprophylaxis with GCS. More than 70 % of the patients had a central venous catheter at inclusion, but less than 10 % in the femoral site. Nearly 30 % of patients had had recent surgery which was predominantly related to the reason for ICU admission. Mean biological values were within normal ranges (Table 1). In 96 of 341 patients for whom the information was available, procoagulant treatment was administered between admission and day 6, 42 in the IPC ? GCS group and 54 in the GCS group (p = 0.12). By day 6, no death was considered by the adjudication committee as due to fatal PE. PE occurred in one patient in the GCS group. No symptomatic DVT was observed. The systematic CUS detected 18 distal DVT and 8 proximal DVT. The incidence of the primary endpoint was 5.6 % (10 of 179 patients) in the IPC ? GCS group

Fig. 1 Flow of patients through the trial

Non included n = 547 - Admission > 36 hours n = 107 - Expected stay < 72 hours n = 141 - Contraindication to GCS and/or IPC n =55 - Life support limitation n = 156 - Refusal n = 53 - Mechanical heart valve n= 14 - Logistical reason n= 13

Screened patients (n = 954)

Randomized patients (n = 407)

IPC + GCS (n = 205)

Excluded n = 1 Nonevaluable for ultrasonography at day6 n = 25 - Death n = 21 - Transfer to another site before ultrasonography n = 3 - Withdrawal of consent n = 1

GCS (n = 202)
Nonevaluable for ultrasonography at day6 n = 18 - Death n = 17 - Transfer to another site before ultrasonography n = 1

Evaluable for efficacy endpoint at day 6 ( n= 179)

Death n = 40

Evaluable for efficacy endpoint at day 6 (n = 184)

Death n = 41

Followed-up until day 30 (n= 142)

Death n = 8 Lost to follow-up n = 1

Followed-up until day 30 (n= 144)

Death n = 10 Lost to follow-up n = 1

Followed-up until day 90 (n = 133)

Followed-up until day 90 (n = 133)

876

and 9.2 % (17 of 184 patients) in the GCS group, leading to a nonsignicant risk reduction of 40.0 % (relative risk, RR, 0.60; 95 % CI 0.281.28; p = 0.19; Table 2). The incidence of VTE was 10.3 % (3 of 29 patients) in patients previously treated with anticoagulant and 6.4 % (24 of 377 patients) in patients without previous anticoagulant treatment (p = 0.41). Among the 96 patients treated with procoagulant treatments between admission and day 6, 7 (7.3 %) had VTE on day 6, and among the 245 patients without procoagulant treatment, 18 (7.3 %) had VTE on day 6 (p = 0.98). The effect of IPC ? GCS on the primary endpoint was globally homogeneous among subgroups (Fig. 2). Anticoagulation was resumed during the ICU stay in
Table 1 Baseline characteristics of patients Characteristics Age (years), mean SD Female gender, n (%) BMI (kg/m2), mean SD SAPS II score, mean SD Primary admission diagnostic category, n (%) Spontaneous intracranial hemorrhage Traumatic intracranial hemorrhage Multisystem trauma Other hemorrhage Severe sepsis or septic shock Acute respiratory distress syndrome with biological abnormalities Other diagnoses On ICU admission, n (%) Hospitalization more than 48 h prior to ICU admission Therapeutic anticoagulation Thromboprophylaxis At inclusion, n (%) Sepsis Central venous catheter Mechanical ventilation Sedation Risk factors for VTE before ICU admission, n Previous VTE Cancer Recent surgery or trauma Pregnancy or post-partum Estrogen use Known thrombophilia Plaster cast immobilization Previous stroke Cardiac insufciency Respiratory insufciency Biological data Platelet count (G/l), mean SD Platelet count \50,000/mm3, n (%) Hemoglobin (g/dl), mean SD Hemoglobin \7 g/dl, n (%) aPTT ratio, mean SD aPTT ratio [2, n (%) Prothrombin time (%), mean SD Prothrombin time \40 %, n (%) Creatinine (lmol/l), mean SD All patients (n = 406) 55.4 17 137 (33.5 %) 25.5 5.2 42.7 18.1 146 (36.0) 87 (21.4) 44 (10.8) 40 (9.9) 39 (9.6) 24 (5.9) 26 (6.4) 68 (16.7) 29 (7.1) 29 (7.1) 46 (11.3) 295 (72.6) 336 (82.8) 317 (78.1) (%) 12 (3.0) 52 (12.8) 118 (29.1) 4 (1.0) 6 (1.5) 2 (0.5) 0 (0) 13 (3.2) 20 (4.9) 64 (15.8) 187.5 108.0 47 (11.6) 12.3 7.5 12 (3.0) 1.16 0.40 8 (2.0) 71.9 21.1 42 (10.3) 101.6 104.7

53 % of survivors, after a median delay of 2 days (interquartile range 027 days) following the day-6 CUS. Between day 6 and day 90, no additional VTE occurred in the GCS group but four symptomatic VTE events occurred in the IPC ? GCS group (two proximal DVT, one distal DVT and one PE). The incidence of symptomatic VTE on day 90 was 2.0 % (4/203 patients) in the IPC ? GCS group and 0.5 % (1/202 patients) in the GCS group (RR 4.0; 95 % CI 0.535.3; p = 0.4). The incidence of all VTE events (symptomatic and asymptomatic) on day 90 was 7.8 % (14/179 patients) in the IPC ? GCS group and 9.2 % (17/184 patients) in the GCS group (RR 0.9; 95 % CI 0.41.7; p = 0.63).

IPC ? GCS group (n = 204) 56.3 16.5 72 (35.3 %) 25.6 4.9 41.6 18.3 81 41 21 19 18 14 (39.7) (20.1) (10.3) (9.3) (8.8) (6.8)

GCS group (n = 202) 54.6 17.5 65 (32.2 %) 25.4 5.5 43.7 17.8 65 46 23 21 21 10 (32.2) (22.8) (11.4) (10.4) (10.4) (5.0)

p value 0.32 0.51 0.72 0.24 0.12 0.51 0.72 0.71 0.59 0.41 0.22 0.83 0.54 0.35 0.33 0.96 0.96 0.95 0.55 0.80 0.78 1.0 0.45 0.15 0.76 0.07 0.96 0.37 0.62 0.08 0.09 0.96 0.49 0.13 0.77 0.12

10 (4.9) 35 (17.2) 13 (6.4) 17 (8.3) 20 (9.8) 148 (72.5) 169 (82.8) 159 (77.9) 5 (2.5) 27 (13.2) 58 (28.4) 2 (1.0) 2 (1.0) 2 (1.0) 0 (0) 6 (3.0) 6 (3.0) 32 (15.7) 192.3 109.2 22 (10.8) 11.6 2.4 9 (4.4) 1.16 0.50 5 (2.5) 73.5 20.5 22 (10.8) 100.6 103.4

16 (7.9) 33 (16.3) 16 (7.9) 12 (5.9) 26 (12.9) 147 (72.8) 167 (82.7) 158 (78.2) 7 (3.4) 25 (12.4) 60 (29.7) 2 (1.0) 4 (2.0) 0 (0) 0 (0) 7 (3.5) 14 (6.9) 32 (15.8) 182.7 107.0 25 (12.4) 12.9 10.4 3 (1.5) 1.15 0.40 3 (1.5) 70.2 21.8 20 (9.9) 102.5 106.2

877 Table 2 Venous thromboembolic events by day 6 Outcome Fatal pulmonary embolism Symptomatic pulmonary embolism Symptomatic deep vein thrombosis Asymptomatic distal deep vein thrombosis Asymptomatic proximal deep vein thrombosis Primary endpoint on day 6 IPC ? GCS group, GCS group, n/N (%) n/N (%) 0/204 0/204 0/204 6/179 (3.4) 4/179 (2.2) 10/179 (5.6) 0/202 1/202 0/202 12/183 (6.6) 4/183 (2.2) 17/184 (9.2) Relative risk 95 % condence interval p value 0.51 1.02 0.60

0.21.33 0.264.03 0.281.28

0.17 0.975 0.191

Compliance with GCS was poor in eight patients (2.0 %, one in the GCS group and seven in the IPC ? GCS group). Tolerance of GCS was poor in 13 patients (3.3 %) in both groups with no difference between groups. The reported adverse effects (none serious) were mild cutaneous injuries (n = 8) or discomfort (sweating, warmth, and tight GCS). Compliance with IPC was poor in 14 patients (7.0 %) in the IPC ? GCS group. Tolerance of IPC was poor in 12 patients (6.0 %) and early removal of IPC was necessary in seven patients. The most frequent reasons for discontinuation of IPC were discomfort, noise and restlessness. No serious adverse event was reported with IPC. By day 6, 15 deaths were related to fatal hemorrhage with no difference between the groups: 7 (3.4 %) in the IPC ? GCS group and 8 (4 %) in the GCS group (p = 0.78). Nonfatal major hemorrhage was observed in 22 patients: 10 (4.9 %) in the IPC ? GCS group and 12 (5.9 %) in the GCS group (p = 0.65). By day 30, 58 patients (29 %) in the GCS group and 61 patients (30 %) in the IPC ? GCS group had died (p = 0.79). By day 90, 68 patients (34 %) in the GCS group and 69 patients (34 %) in the IPC ? GCS group had died (p = 0.97).

Discussion
In this rst randomized trial specically designed to evaluate the potential benet of two distinct mechanical devices in ICU patients with a high risk of bleeding, we found no signicant difference in the incidence of VTE between those receiving IPC together with GCS and those receiving GCS alone. For critically ill patients who are at high risk of bleeding, current guidelines recommend the optimal use of mechanical thromboprophylaxis with GCS and/or IPC to prevent VTE at least until the bleeding risk decreases [17]. However, little is known about the efcacy and safety of mechanical prophylaxis methods in ICU patients [21]. From a physiological point of view, the dynamic compression produced by IPC is assumed to be more

efcient than the static compression produced by the GCS. In addition, IPC is thought to have a brinolytic action, primarily by increasing tissue plasminogen activator and decreasing plasminogen activator inhibitor [22]. Despite these theoretical advantages, the addition of IPC provided no additional benet compared to GCS alone in our ICU patients with a high risk of bleeding. Nevertheless, the lack of statistical power of the present study (54 % when calculated a posteriori) did not allow us to condently rule out a potential benecial effect of IPC for the prevention of VTE in this specic clinical setting. This low statistical power may be explained by the low incidence of VTE observed in the GCS group (9.2 %), when compared to an expected 15 % rate. In a previous randomized clinical trial that evaluated IPC ? GCS versus GCS alone in patients with intracerebral hemorrhage, we found a 15.9 % incidence of VTE in the GCS group using systematic CUS on day 10 [19]. In the present trial, the incidence of VTE in a similar population subset (patients with intracranial hemorrhage in the GCS group, n = 103) was 12.6 %. In an observational study performed in 100 medical ICU patients, lower limb Doppler ultrasonography performed twiceweekly and 1 week after ICU discharge detected a DVT in 33 % of patients who were receiving a mechanical prophylaxis [6]. A prospective study using twice-weekly leg ultrasonography identied a proximal DVT in 25 of 261 patients (10 %) during their ICU stay [3]. Taken together, these results led us to initially expect a 15 % incidence of VTE in the GCS group, including asymptomatic distal and proximal DVT. The lower incidence of VTE observed in the control group of the present study may be related to a lower risk of thrombosis in critically ill patients at high risk of bleeding when compared to that of the general ICU population. Early CUS evaluation on day 6 presumably contributed to the fairly low incidence of observed VTE in the present study. In previous trials, VTE endpoints were generally evaluated on day 10 or later. Nevertheless, in studies that screened regularly for DVT in ICU patients, DVT were identied in the rst days of ICU stay. In a study by Hirsch et al., 23 of 33 DVT (70 %) were detected during the rst CUS [6].

878

Fig. 2 Rates of the primary efcacy outcome on day 6 in subgroups. The primary efcacy endpoint was a composite, combining fatal PE, symptomatic PE, symptomatic DVT, or asymptomatic DVT detected by systematic CUS on day 6. The size of each square is in proportion to the number of patients in the comparison. No adjustment for multiple comparisons was made, since subgroup analyses were performed for exploratory purposes only. BMI is body weight in kilograms divided by the square of the height in meters

Similarly, Cook et al. [3] reported that 50 % of DVT were detected before day 8 (median day 8; interquartile range day 414). We also chose to perform the CUS evaluation early because in patients admitted to an ICU with a high bleeding risk, the American College of Chest Physicians guidelines recommend that

pharmacological thromboprophylaxis be substituted for or added to the mechanical thromboprophylaxis as soon as the high bleeding risk decreases [17]. In the present study, anticoagulation was resumed during the ICU stay in the majority of the survivors early after the systematic CUS on day 6.

879

Since VTE is often silent and clinically unsuspected in ICU patients [10, 11], PE is one of the most common unexpected autopsy ndings in this specic population [11]. The recent PROTECT study showed a signicantly reduced incidence of PE in ICU patients treated with dalteparin when compared with the group treated with unfractionated heparin, but no reduction in asymptomatic DVT [16]. This result questions the validity of asymptomatic DVT as a surrogate primary endpoint in prophylaxis trials in VTE. In the present trial, we failed to detect any intergroup difference in clinically relevant events, such as symptomatic VTE or death. The compliance with IPC was fairly good in the present study. The high proportion of sedated patients at inclusion (78 %) presumably accounts for this result. Nevertheless, a few patients without sedation poorly tolerated the ination or the noise of IPC. No serious adverse event was related to the use of IPC. The present trial had substantial limitations. First, as discussed above, the study was underpowered to demonstrate a signicant 60 % reduction in incidence in the IPC ? GCS group. Second, autopsy was not performed for all fatalities, especially when death occurred in the ICU. This may have led to underestimation of the PErelated death rate. Third, this study was not blinded, but day 6 CUS was performed without knowledge of the mechanical device used for VTE prophylaxis and validation of CUS was performed by an expert ultrasonographer who was unaware of the patients assignment. In conclusion, the current trial was not able to show the superiority of the combination of IPC ? GCS over GCS alone in preventing VTE in ICU patients at high risk of bleeding. However, an additional preventing effect of IPC cannot be condently ruled out based on the results of this study which lacked power, mainly because of a low incidence of VTE in the control group. Nevertheless, this study provides the basis for performing new large clinical trials to evaluate mechanical devices to prevent VTE in various subsets of ICU patients with a high risk of bleeding.
Acknowledgments This study was supported by a grant from the French Ministry of Health (PHRCN 2005 no. 08-13). COVIDIEN supplied the GCS (T.E.D.TM anti-embolism stockings) and the IPC (SCD EXPRESSTM compression systems with tubing sets and sleeves). The sponsor had no role in the study. We thank the research coordinator I. Pichon, the members of the CIC-P 0502 and the CIC-P 0801, the study nurses of the Clinical Research in Intensive Care and Sepsis Group (CRICS group), the Direction de la Recherche Clinique et Innovation (DRCI) of Brest University Hospital for their invaluable work, as well as Z. Alavi for her pertinent advice. Conicts of interest The authors state that they have no conict of interest.

Appendix
CIREA1 collaborators: Steering Committee: Pr Jean-Luc Diehl, Pr Philippe Vignon, Dr Anne Renault, Pr Karine Lacut Adjudication Committee: Pr L. Bressollette (expert ultrasonographer), Pr Erwan LHer Data manager and Biostatistician: Elise Poulhazan, Emmanuel Nowak Investigators: animation Polyvalente, CH dAngoule me, Service de Re Saint-Michel: Dr Thierry Baudin-Jacquemin, Dr Sylvie Nicole Calvat, Dr Arnaud Desachy, Dr Florence Hospital animation Me dicale, CHU de la Cavale Service de Re ` vre, Dr Jean-Marie Blanche, Brest: Dr Montaine Lefe Tonnelier, Dr Alexandre Tonnelier de ration dAnesthe sie, Re animation Urgences, HIA Fe Clermont-Tonnerre, Brest : Dr Christophe Giarcardi, Dr Bruno Ralec, Dr Mehdi Ould-Ahmed, Dr Didier Fourel, Dr Ba Vinh N Guyen, Dr Diane Commandeur. animation Polyvalente, CHU de Limoges, Limoges: Re Dr Jean-Bernard Amiel, Dr Marc Clavel, Dr Anthony Dugard, Dr Caroline Etchecopar-Chevreuil, Dr Bruno Franc ois, Dr Nicolas Pichon, Dr Jean-Claude Voultoury. animation Me dicale, CHU Ho tel Dieu, Service de Re Nantes: Dr Olivier Zambon animation, CHRHo pital de la Source, Orle ans: Re Dr Isabelle Runge, Dr Christian Fleury, Dr Marie Skarzynski, Dr Dalila Benzekri-Lefevre, Dr Anne Bretagnol, Dr Nicolas Bercault animation Me dicale, HEGP, AP-HP Paris: Dr EmRe manuel Guerot, Dr Ana Novara, Dr Christophe Faisy, Dr Iris Pelieu, Pr Nicolas Lerolle animation Me dicale, CHU de Poitiers, Service de Re Poitiers: Pr Olivier Pourrat, Dr Julien Voultoury, Dr Michel Pinsard, Dr Anne Veinstein, Dr Jean-Pierre Frat, Robert Dr Julie Badin, Pr Rene animation Me dicale, CHU de Tours Service de Re pital Bretonneau, Tours: Dr Annick Legras, Dr EmHo phane manuelle Mercier, Dr Laure Batias-Moreau, Dr Ste phanie Benardeau, Dr Maud Jonas, Ehrmann, Dr Ste Dr Antoine Guillon, Dr Jennifer Buret, Dr Elodie Masseret Vascular Ultrasonography Collaborators: dicale, CH dAngoule me, Saint-Michel: Imagerie Me ` le Colin, Dr Thierry Landois Dr Florence Hospital, Dr Danie Radiologie, HIA, Clermont-Tonnerre, Brest : Dr Jean Rousset, Dr Marc Garetier, Dr Diouf, Dr Sandra Chinellato, Dr Valentin Tissot Vasculaire du service de Chirurgie Thoracique Unite et Cardiovasculaire, CHU de Limoges, Limoges: Pr Philippe Lacroix, Dr Tiphaine Bonnafy tel Dieu, Service dexplorations vasculaires, CHU Ho ro me Connault Nantes: Dr Je

880

pital de la Source, Orle ans: Radiologie, Echographie-Doppler, CHU de Poitiers, Angiologie, CHRHo cile Thollot, Dr Je rPoitiers: Dr Elisabeth Escure, Dr Ce Dr Carole Bazzi me Roumy decine vasculaire et hypertension arte rielle, o Me le imagerie me dicale, CHU de ToursHo pital Po HEGP, AP-HP Paris: Pr Joseph Emmerich, Dr Emman de ric Patat Bretonneau, Tours: Dr Nicole Ferreira, Pr Fre uel Messas

References
1. Attia J, Ray JG, Cook DJ, Douketis J, Ginsberg JS, Geerts WH (2001) Deep vein thrombosis and its prevention in critically ill adults. Arch Intern Med 161:12681279 2. Shorr AF, Williams MD (2009) Venous thromboembolism in critically ill patients. Observations from a randomized trial in sepsis. Thromb Haemost 101:139144 3. Cook D, Crowther M, Meade M, Rabbat C, Grifth L, Schiff D, Geerts W, Guyatt G (2005) Deep venous thrombosis in medical-surgical critically ill patients: prevalence, incidence, and risk factors. Crit Care Med 33:15651571 4. Cook D, Douketis J, Meade M, Guyatt G, Zytaruk N, Granton J, Skrobik Y, Albert M, Fowler R, Hebert P, Pagliarello G, Friedrich J, Freitag A, Karachi T, Rabbat C, Heels-Ansdell D, Geerts W, Crowther M; Canadian Critical Care Trials Group (2008) Venous thromboembolism and bleeding in critically ill patients with severe renal insufciency receiving dalteparin thromboprophylaxis: prevalence, incidence and risk factors. Crit Care 12:R32 5. Geerts WH, Heit JA, Clagett GP, Pineo GF, Colwell CW, Anderson FA Jr, Wheeler HB (2001) Prevention of venous thromboembolism. Chest 119:132S175S 6. Hirsch DR, Ingenito EP, Goldhaber SZ (1995) Prevalence of deep venous thrombosis among patients in medical intensive care. JAMA 274:335337 7. McKelvie PA (1994) Autopsy evidence of pulmonary thromboembolism. Med J Aust 160:127128 8. Moser KM, Fedullo PF, LitteJohn JK, Crawford R (1994) Frequent asymptomatic pulmonary embolism in patients with deep venous thrombosis. JAMA 271:223225 9. Stein PD, Henry JW (1995) Prevalence of acute pulmonary embolism among patients in a general hospital and at autopsy. Chest 108:978981 10. Karwinski B, Svendsen E (1989) Comparison of clinical and postmortem diagnosis of pulmonary embolism. J Clin Pathol 42:135139 11. Twigg SJ, McCrirrick A, Sanderson PM (2001) A comparison of post mortem ndings with post hoc estimated clinical diagnoses of patients who die in a United Kingdom intensive care unit. Intensive Care Med 27:706710 12. Blosser SA, Zimmerman HE, Stauffer JL (1998) Do autopsies of critically ill patients reveal important ndings that were clinically undetected? Crit Care Med 26:13321336 13. Mort TC, Yeston NS (1999) The relationship of pre-mortem diagnoses and post-mortem ndings in a surgical intensive care unit. Crit Care Med 27:299303 14. Cade JF (1982) High risk of the critically ill for venous thromboembolism. Crit Care Med 10:448450 15. Fraisse F, Holzapfel L, Couland JM, Simonneau G, Bedock B, Feissel M, Herbecq P, Pordes R, Poussel JF, Roux L (2000) Nadroparin in the prevention of deep vein thrombosis in acute decompensated COPD. Am J Respir Crit Care Med 161:11091114 16. PROTECT Investigators for the Canadian Critical Care Trials Group and the Australian and New Zealand Intensive Care Society Clinical Trials Group, Cook D, Meade M, Guyatt G, Walter S, Heels-Ansdell D, Warkentin TE, Zytaruk N, Crowther M, Geerts W, Cooper DJ, Vallance S, Qushmaq I, Rocha M, Berwanger O, Vlahakis NE (2011) Dalteparin versus unfractionated heparin in critically ill patients. N Engl J Med 364:13051314 17. Kahn SR, Lim W, Dunn AS, Cushman M, Dentali F, Akl EA, Cook DJ, Balekian AA, Klein RC, Le H, Schulman S, Murad MH; American College of Chest Physicians (2012) Prevention of VTE in nonsurgical patients: Antithrombotic Therapy and Prevention of Thrombosis, 9th ed: American College of Chest Physicians EvidenceBased Clinical Practice Guidelines. Chest 141(2 Suppl):e195Se226S 18. Turpie AGG, Hirsh J, Gent M, Julian D, Johnson J (1989) Prevention of deep vein thrombosis in potential neurosurgical patients. A randomized trial comparing graduated compression stockings alone or graduated compression stockings plus intermittent pneumatic compression with control. Arch Intern Med 149:679681 19. Lacut K, Bressollette L, Le Gal G, Etienne E, De Tinteniac A, Renault A, Rouhart F, Besson G, Garcia JF, Mottier D, Oger E, VICTORIAh (Venous Intermittent Compression and Thrombosis Occurrence Related to Intra-cerebral Acute hemorrhage) Investigators (2005) Prevention of venous thrombosis in patients with acute intra-cerebral hemorrhage. Neurology 65:865869 20. Goodacre S, Sampson F, Thomas S, van Beek E, Sutton A (2005) Systematic review and meta-analysis of the diagnostic accuracy of ultrasonography for deep vein thrombosis. BMC Med Imaging 5:6 21. Limpus A, Chaboyer W, McDonald E, Thalib L (2006) Mechanical thromboprophylaxis in critically ill patients: a systematic review and metaanalysis. Am J Crit Care 15:402410 22. Jacobs DG, Piotrowski JJ, Hoppensteadt DA, Salvator AE, Fareed J (1996) Hemodynamic and brinolytic consequences of intermittent pneumatic compression: preliminary results. J Trauma 40:710716