Endovascular cooling versus standard femoral catheters and intravascular complications: A propensity-matched cohort study

Resuscitation 124 (2018) 1–6

Contents lists available at ScienceDirect

Resuscitation journal homepage: www.elsevier.com/locate/resuscitation

Clinical paper

Endovascular cooling versus standard femoral catheters and intravascular complications: A propensity-matched cohort study夽 Olivier Andremont a,∗ , Damien du Cheyron a , Nicolas Terzi a,b,c,d , Cedric Daubin a , Amélie Seguin a , Xavier Valette a , Flore-Anne Lecoq e , Jean-Jacques Parienti e,f,1 , Bertrand Sauneuf a,g,1 a

Service de Réanimation médicale, Centre Hospitalier Universitaire de Caen, avenue de la cote de Nacre, 14033 Caen, France Inserm U 1075 COMETE, 14032, Caen, France c Service de réanimation médicale, CHU Grenoble Alpes, 38000, Grenoble, France d HP2, Inserm U1042, Université Grenoble-Alpes, 38000, Grenoble, France e Unité de biostatistique et recherche Clinique, Centre Hospitalier Universitaire de Caen, avenue de la cote de Nacre, 14033 Caen, France f EA2656 Groupe de Recherche sur l’Adaptation Microbienne (GRAM 2.0), Université Caen Normandie, France g Service de Réanimation Médicale Polyvalente, Centre Hospitalier Public du Cotentin, BP 208, 50102 Cherbourg-en-Cotentin, France b

a r t i c l e

i n f o

Article history: Received 27 July 2017 Received in revised form 7 December 2017 Accepted 11 December 2017 Keywords: Cardiac arrest Therapeutic hypothermia Targeted temperature management Central venous catheter Cooling catheter Catheter-related-thrombosis

a b s t r a c t Background: Targeted temperature management (TTM) contributes to improved neurological outcome in adults who have been successfully resuscitated after cardiac arrest with shockable rhythm. Endovascular cooling catheters are widely used to induce and maintain targeted temperature in the ICU. The aim of the study was to compare the risk of complications with cooling catheters and standard central venous catheters. Materials and methods: In this prospective single-centre cohort study, we included all patients admitted to an intensive care unit for successfully resuscitated cardiac arrest that required endovascular TTM ® (Coolgard , ZollTM Medical corporation, MA, USA), between August 2012 and November 2014, inclusive. We matched the endovascular cooling catheter cohort with a retrospective historical cohort of 512 central femoral venous catheters from the 3SITES trial to compare thrombotic and infectious complications. Results: Overall, 108 patients were included in the cooling cohort, of which 89 had ultrasound doppler. The duration of catheterization was 4.9 days in the control group versus 4.2 days in the TTM group (p = 0.08). After propensity-score matching, there were significantly more thrombotic complications in the cooling (n = 75) than in the control (n = 75) group (12 of 75 (16%) versus 0 of 75 (0%), respectively, p = 0.005), and 4 patients presented major complications. There were 8 colonized catheters in each group (11%) (p > 0.99), and none of the patients had a catheter-related bloodstream infection. Conclusions: In our propensity-score matched study, endovascular cooling catheters were associated with an increased risk of venous catheter-related thrombosis compared to standard central venous catheters. © 2017 Published by Elsevier Ireland Ltd.

Introduction Since the early 2000s, targeted temperature management (TTM) (32–34 ◦ C) has been shown to improve outcomes in adults successfully resuscitated after cardiac arrest with shockable rhythm.

夽 A Spanish translated version of the abstract of this article appears as Appendixi n the final online version at https://doi.org/10.1016/j.resuscitation.2017.12.014. ∗ Corresponding author. Current address: Service de Réanimation médicale, Centre Hospitalier Universitaire de Caen, avenue de la cote de Nacre, 14033 Caen, France. E-mail address: [email protected] (O. Andremont). 1 Dr Sauneuf and Parienti contributed equally to this work. https://doi.org/10.1016/j.resuscitation.2017.12.014 0300-9572/© 2017 Published by Elsevier Ireland Ltd.

Recently, Nielsen et al. has clarified this finding by demonstrating that a target of 36◦ gives similar results [1]. Consequently, recent guidelines underline TTM as a key element in the management of resuscitated out-of-hospital cardiac arrest (OHCA) [1–3]. Different cooling techniques and cooling devices have been developed to maintain optimal TTM. Intravenous infusion of cold fluid has been proposed [4]. External cooling by ice pack or blanket are also widely used [5]. An endovascular method with a dedicated cooling catheter has also been developed, allowing a better control of time to target temperature and better control of temperature during the maintenance phase [6–8]. Additionally, the intravascular method results in fewer variations than skin conductivity by the external

2

O. Andremont et al. / Resuscitation 124 (2018) 1–6

Fig. 1. Flow chart of the study.

cooling method [9]. Despite these advantages, the intravascular method has not demonstrated a clear advantage in terms of prognosis in comparison with the external method [10–12]. Moreover, some side effects related to this cooling method have been reported [13,14]. In fact, infections are frequent after cardiac arrest and may be even more so after TTM [15]. Catheter-related thrombosis is also associated with the use of intravascular devices [16,17], particularly when inserted into the jugular or femoral veins [18], and the majority of these are asymptomatic. Nevertheless, all deep-vein thromboses can lead to pulmonary embolism [19]. The activation of blood coagulation and fibrin formation, which is insufficiently balanced by fibrinolysis among patients with cardiac arrest, probably increases the risk of clot formation [20,21]. Additionally, risks of infection and venous thrombosis are often associated with one another [16]. Post-cardiac arrest patients frequently need central venous access. It may therefore be convenient to use an endovascular cooling catheter, as both a cooling device and central venous access. However, the risk associated with this practice has not been assessed. To fill part of this gap, we performed a prospective monocentric study to evaluate the possible additional side effects associated with the use of endovascular cooling catheters in comparison to standard central venous catheters. Methods Study design and patients This prospective observational study was performed in the medical intensive care unit of the university hospital of Caen, France, between August 2012 and November 2014, inclusive. The study protocol was approved by the French national ethical review board (CPP nord-ouest III Ref: A13-D27-VOL.17) and was con-

ducted according to the principles of the Declaration of Helsinki. We included all adult patients admitted to the intensive care unit (ICU) requiring TTM, as judged by the physician in charge, and in whom an endovascular cooling catheter was used (EvalCool database). Exclusion criteria were pregnancy, non-French speaking patients or families, or those protected by law. All data were collected by the principal investigator of the study. Patients were followed until discharged from the ICU. Initial informed consent was obtained from families and confirmed by patients with favourable neurological outcome. To compare endovascular cooling catheters to standard central venous catheters, we matched patients with cooling catheters with control patients from a previously published trial (3SITES database [18]) who had femoral central venous catheters. The 3SITES study included 3471 standard central venous catheters, and the primary outcome was a composite of catheter-related bloodstream infection and symptomatic deep-vein thrombosis. Procedures The femoral site was used for insertion of all cooling catheters. ® ® In all instances, Coolgard catheters (Coolgard , ZollTM Medical Corporation, MA, USA) were used, and thus, only the femoral catheters from the control cohort were used for comparison. The ® Coolgard device circulates water in a closed system through an endovascular catheter with water temperature controlled by way of “closed-loop” temperature feedback from the patient. All procedures were performed or supervised by highly trained residents or staff physicians who had performed at least 50 previous femoral catheterizations. Surgical hand antisepsis, sterile gloves, masks, surgical gowns, and caps were always used. Patients were covered by sterile drapes from the head to feet. Choices of antiseptic and ultrasonographic guidance or anatomic landmarks for catheter insertion were left to the judgement of the physician. Management

O. Andremont et al. / Resuscitation 124 (2018) 1–6

3

Table 1 Patient baseline characteristics. a: unmatched, b: matched. a: Unmatched Baseline characteristics unmatched

Cohort n = 601

Coolgard n = 89

3sites n = 512

p-value

Age, mean (SD) Male, n (%) SOFA, mean (SD) IGSII, mean (SD) Duration of catheterization, mean (SD) Antibiotherapy, n (%) Anticoagulation, n (%)

60.6 (15.1) 400 (66.6) 8.4 (4.5) 55.7 (18.0) 5.6 (4.1) 316 (52.6) 241 (40.1)

60.7 (15.6) 73 (82.0) 10.9 (6.1) 54.0 (17.8) 4.0 (2.1) 21 (23.6) 65 (73.0)

59.7 (12.7) 327 (63.9) 8.1 (4.0) 65.3 (16.5) 5.9 (4.3) 295 (57.6) 176 (34.4)

0.51 0.0006 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001

Baseline characteristics matched

Cohort n = 150

Coolgard n = 75

3 sites n = 75

p-value

Age, mean (SD) Male, n (%) SOFA, mean (SD) IGSII, mean (SD) Duration of catheterization, mean (SD) Antibiotherapy, n (%) Anticoagulation, n (%)

59.8 (13.9) 119 (0.79) 10.3 (3.3) 64.1 (15.4) 4.5 (2.7) 45 (0.30) 109 (0.73)

59.3 (12.9) 60 (0.80) 10.0 (3.2) 64.1 (14.5) 4.2 (2.1) 21 (0.28) 52 (0.69)

60.2 (14.9) 59 (0.79) 10.5 (3.3) 64.2 (16.4) 4.9 (3.1) 24 (0.32) 57 (0.76)

0.71 1.0 0.30 0.94 0.08 0.70 0.36

b: Matched

of target temperature of 33–34 ◦ C was induced and maintained for 12–24 h. Decisions to remove the catheters were made by the physician in charge of the patient. All catheters tips were cultured. Thrombosis was systematically searched for by ultrasonography of the femoral axis and the inferior vena cava between 24 h and 48 h after removal of the catheter. Measures and outcomes Catheter-related colonization was defined by a culture of the ˆ catheter tip ≥ 103 UFC. Thrombosis was defined by the presence of a thrombus upon ultrasonography at the catheter vascular insertion site or in the inferior vena cava. If pulmonary embolism was suspected, a chest CT scan was performed. The primary outcome was the incidence of cooling catheterrelated complications from the time of catheter insertion to 48 h after catheter removal (catheter colonization and catheter related thrombosis). Major complications were either catheter-related bloodstream infection or catheter-related symptomatic deep-vein thrombosis. Catheter-related bloodstream infection was defined as the presence of bacteraemia originating from an intravenous catheter (refers to same microbial pathogen in blood culture and quantitative culture of the catheter tip) (The Medical Dictionary for Regulatory Activities (MedDRA), version 17, code 10064687, grade 3 or higher); catheter-related symptomatic deep vein thrombosis was defined by ultrasonography-confirmed thrombosis at the catheter insertion site, associated with clinical symptoms (i.e., swelling, pain) (MedDRA, version 17, code 10062169, grade 3 or higher) [22]. Statistical analysis No a priori sample size calculation was computed for this study, and we planned empirically to include 100 patients in the EvalCool cohort. The data are expressed as the mean ± standard deviations (SD) or median (interquartile range) and percentage, depending on the nature of the variable of interest. Since the study was not randomized, the patients included in the EvalCool and 3SITES cohorts could not have the same baseline risk of catheter thrombosis or infection. Therefore, potential indication or “channelling” biases were adjusted for by developing a propensity score for using one

versus another catheter. A stepwise logistic regression analysis was performed to select baseline variables, including first-order interactions that were associated with the use of cooling catheters. Variables were entered into the model at a P value cut-off of 0.50 to derive a full non-parsimonious model. Using these selected variables, a propensity score was estimated by maximum likelihood logistic regression analysis. Calibration of the final logistic model was assessed using the Hosmer-Lemeshow statistic. We used 1:1 greedy propensity-score matching (PSM) to ensure that the cooling catheter and standard central venous catheter populations were comparable at baseline. In unadjusted analyses, cooling and standard catheters were compared with the use of a chi-square test or Student or Wilcoxon tests, as appropriate. After PSM, the 2 groups were compared with the use of paired tests, such as the McNemar chi-square or the t-test for paired data, as appropriate. A P value < 0.05 was considered to be significant, and all P values were two-tailed without adjusting for multiple comparisons. The statistical analyses were performed using SAS statistical software, version 9.4 (SAS Institute Inc., Cary, NC). This was the first analysis of the data since no interim analysis was planned or conducted. Results Patient characteristics One hundred and eight patients with cooling venous catheters (EvalCool database) were assessed. All had been admitted to ICU after resuscitated cardiac arrest. Nineteen were excluded because no doppler ultrasonography was performed after the catheter was removed. In the 3SITES database, data for only 512 patients were used, the others having either a non-femoral catheter or no doppler result (Fig. 1). Baseline characteristics of the patients before matching are shown in Table 1. The median age was 60.6 (SD 15.1) years with no significant differences between groups (p = 0.51). In contrast, there were significantly more males in the cooling cohort than in the controls (73 (82.0%) versus 327 (63.9%), p = 0.0006). Severity scores of cooling patients were higher (10.9 (SD 6.1) versus 8.1 (SD 4.0) for the SOFA score, p < 0.0001; and 55.7 (SD 18.0) versus 54.0 (SD 17.8) for the IGSII, p < 0.0001). The length of catheterization was longer in the controls (5.6 (SD 4.1) days versus 4.0 (SD 2.1) days, p < 0.0001), and patients from the 3SITES cohort

4

O. Andremont et al. / Resuscitation 124 (2018) 1–6

Table 2 Outcomes. a: unmatched. b: ps-matched (ps; propensity score). a: ps-Unmatched Outcome unmatched

Cohort n = 601

Coolgard n = 89

3 sites n = 512

p-value

Colonization, n (%) Thrombosis, n (%)

86 (14) 36 (6)

9 (10) 15 (17)

77 (15) 21 (4)

0.25 <0.0001

Outcome ps-matched

Cohort n = 150

Coolgard n = 75

3 sites n = 75

p-value

Colonization, n (%) Thrombosis, n (%)

16 (11) 12 (8)

8 (11) 12 (16)

8 (11) 0 (0)

>0.99 0.005

b: Matched

received more antibiotics (57.6% versus 23.6%, p < 0.0001). Patients from the cooling cohort received more anticoagulants (73% versus 34.4%, p < 0.0001), probably because they had all been resuscitated from cardiac arrest (Table 1a). After matching, there were no further significant differences in the characteristics of the two groups (Table 1b). After propensity score matching, 75 patients from each group were analysed (Fig. 1). Major complications in the EvalCool cohort In the cooling cohort, 4 major complications occurred. In one patient, haemorrhagic shock complicated catheter insertion (grade 4). Three patients suffered from symptomatic pulmonary emboli; of these, one survived following severe acute respiratory distress (grade 4), one died of refractory cardiogenic shock (grade 5), and another died after cerebral haemorrhage secondary to intravenous thrombolysis (grade 5). None of the patients had catheter-related bloodstream infections. Overall outcomes and unadjusted analyses In the unadjusted analysis, patients with cooling catheters had more thrombotic complications than the control patients (15 of 89 (17%) versus 21 of 512 (4%), p < 0.0001), although they had received anticoagulants more often (p < 0.0001). Colonization rates of catheter tips were not significantly different between the two groups, (9 of 89 (10%) versus 77 of 512 (15%) in the cooling versus control patients, respectively, p = 0.25) (Table 2a). Propensity score-matched analysis There were more thrombotic complications in the cooling group than in the control group (respectively, 12 of 75 (16%) versus 0 of 75 (0%), p = 0.005). There were 8 catheters colonized in each group of 75 (11%) (p > 0.99) (Table 2b). Discussion In this prospective observational matched-cohort study, endovascular cooling catheters were associated with a significantly increased rate of thrombosis compared to control standard venous catheters. Additionally, we reported major thrombotic events such as pulmonary embolism. To our knowledge, our study was the first prospective study comparing thrombotic complications between cooling and standard catheters in the setting of cardiac arrest. An increased risk of thrombosis was reported using the same device in a Canadian

study. However, this risk seemed to be abrogated by the nonrandomized use of prophylactic anticoagulation [13]. We did not confirm this result. In our unadjusted analyses, most of the EvalCool patients received anticoagulants, and there were more thromboses. Therefore, the increased risk of thrombosis with an endovascular cooling catheter was not linked to the anticoagulant treatment in our study. Subarachnoid haemorrhage is a situation in which TTM is used [23]. Studies have also reported frequent thrombosis associated with the use of intravascular cooling catheters in this setting [14]. Others have used invasive cavography before removal of the cooling catheter and found a higher positive predictive value compared to ultrasound for the diagnosis of intravascular thrombosis. In these studies, the rate of catheter-related thrombosis could reach 50%-90% [24,25]. Consequently, our rate of thrombosis may be underestimated. Our ultrasonography doppler examination only investigated the femoral vein and the lower segment of the inferior vena cava. Finally, cardiac arrest itself did not appear as a risk factor for thrombosis in two studies investigating coagulation during TTM after OHCA [26,27]. Many studies have demonstrated an increased risk of infections after TTM [15,28]. These findings are supported by experimental data showing depressed immune responses during the hypothermia phase [29]. Nearly two-thirds of patients receiving TTM at 33–34 ◦ C appeared to develop infections during their ICU stay [15], irrespective of whether temperature was maintained between 33 ◦ C and 36 ◦ C [30]. Most often, the infected sites were lungs. Central venous line-related infection may also occur more frequently [15]. In our study, we did not find any difference in terms of colonization, which is considered to accurately estimate the risk of infection [31]. This is in line with a previous single-centre study that reported no cooling catheter-related infection [32]. Whether the presence of this specific device may have led to specific care influencing colonization in comparison with standard central venous catheters remains unknown. Hence, our data does not argue that an increased risk of infection with the intravascular cooling catheter exists. Another hypothesis is the potential influence of hypothermia on the risk of vascular thrombosis. The results concerning the risk of arterial stent thrombosis during TTM are controversial. Large studies did not find any increase in the incidence of stent thrombosis in OHCA patients undergoing percutaneous coronary intervention [33]. Others described a trend of higher risk for stent thrombosis, a risk that could be counter-balanced by antithrombotic treatment [34]. Concerning the risk of deep vein thrombosis during TTM, few data are available. The recent Epo-ACR-02 study reported a deep vein thrombosis rate of 5.8% in the non-erythropoietin-treated group [35]. In fact, an ancillary work of the recent TTM study [1] did not find substantial differences in coagulation between TTM 33–34 ◦ C versus 36 ◦ C in out-of-hospital cardiac arrest patients, raising the question of the true impact of hypothermia on the risk of vascular thrombosis [27]. If the respective role of cardiac arrest itself and hypothermia as a risk factor for thrombosis remains unclear, our main hypothesis is that the shape of the catheter, its length and its diameter, which is larger than that of standard central venous catheters, could decrease venous flow and promote blood stasis. As a consequence, it could increase the risk of clot formation and explain the higher rate of catheter-related thrombosis we observed. Several studies have compared endovascular versus external TTM after OHCA. None of the techniques confer an advantage in terms of survival without major neurological damage. Theoretically, endovascular cooling devices confer better temperature control [10–12,36], which leads to faster cooling, an improved thermic stability and better temperature control during the rewarming phase, avoiding overcooling and rebound hyperthermia [8,9]. Side effects of endovascular cooling catheters have been inconsistently

O. Andremont et al. / Resuscitation 124 (2018) 1–6

reported in these studies. None of the studies reported severe complications related to venous thrombosis. However, doppler examination was not systematically performed, and only symptomatic catheter related-thrombosis was reported. Here, we also report major complications including two deaths resulting directly from consequences of the thrombosis or from the anticoagulant treatment. It is surprising that such severe side effects had not been reported previously. We hypothesized that a number of pulmonary emboli were not diagnosed and attributed to another complication, in the context of post-cardiac arrest shock. Our study has several limitations. First, an important difference between the two studied groups was the temperature. Cardiac arrest patients were managed with TTM at 33–34 ◦ C, so it is difficult to conclude the respective role of the catheter device and of the hypothermia in the genesis of thrombosis. Second, we evaluated the catheter-associated complications after a catheterization time that is longer than the recommended TTM period. We acknowledge that this can be seen as bias. However, we think this represents the usual practice with these catheters. Indeed, the presence of infusion lumens may avoid the placement of another standard venous catheter, and cooling catheters are often not removed at the end of the TTM period. In the cooling group, we only included patients with cardiac arrest representing a specific condition. Although compared patients were severity-matched based on IGSII, biological abnormalities presenting after cardiac arrest may have promoted thrombosis. Our findings on the colonization of each catheter type may have been biased by the absence of skin antisepsis control. Finally, although we used sophisticated statistical methods to limit bias, our research question cannot be answered by a randomized trial, and therefore, residual confounding may persist. Conclusion The use of endovascular cooling catheters was associated with a higher risk of venous catheter-related thrombosis compared with standard central venous catheters, with potential severe complications, including death. Given the lack of survival benefit provided by this device [12,36], the risk-benefit ratio and costs of using systematic endovascular cooling catheters instead of other methods should therefore be carefully appraised. Additionally, endovascular cooling catheters should be promptly removed at the end of the TTM period, and a new standard central venous catheter inserted if central venous access is still required, ideally in the subclavian vein [18]. Whether the risk of thrombosis depends only on the catheter device or on the biological changes induced by hypothermia and cardiac arrest remains to be investigated. Conflict of interest The authors declare that they have no competing interests. Acknowledgements This work was supported by institutional and departmental funds (Caen Teaching Hospital and Intensive Care Unit, respectively). References [1]. Nielsen N, Wetterslev J, Cronberg T, Erlinge D, Gasche Y, Hassager C, et al. Targeted temperature management at 33 ◦ C versus 36 ◦ C after cardiac arrest. N Engl J Med 2013;369(23):2197–206. [2]. Hypothermia after Cardiac Arrest Study Group. Mild therapeutic hypothermia to improve the neurologic outcome after cardiac arrest. N Engl J Med 2002;346(8):549–56.

5

[3]. Hazinski MF, Nolan JP, Aickin R, Bhanji F, Billi JE, Callaway CW, et al. Part 1: executive summary. International consensus on cardiopulmonary resuscitation and emergency cardiovascular care science with treatment recommendations. Circulation 2015;132(16 Suppl. 1):S2–39. [4]. Bruel C, Parienti J-J, Marie W, Arrot X, Daubin C, Du Cheyron D, et al. Mild hypothermia during advanced life support: a preliminary study in out-ofhospital cardiac arrest. Crit Care 2008;12(1):R31. [5]. Deye N, Vincent F, Michel P, Ehrmann S, da Silva D, Piagnerelli M, et al. Changes in cardiac arrest patients’ temperature management after the 2013 TTM trial: results from an international survey. Ann Intensive Care 2016;6(1):4. [6]. Hoedemaekers CW, Ezzahti M, Gerritsen A, van der Hoeven JG. Comparison of cooling methods to induce and maintain normo- and hypothermia in intensive care unit patients: a prospective intervention study. Crit Care 2007;11(4):R91. [7]. Polderman KH, Herold I. Therapeutic hypothermia and controlled normothermia in the intensive care unit: practical considerations, side effects, and cooling methods. Crit Care Med 2009;37(3):1101–20. [8]. Knapik P, Rychlik W, Siedy J, Nadziakiewicz P, Cie´sla D. Comparison of intravascular and conventional hypothermia after cardiac arrest. Kardiol Pol 2011;69(11):1157–63. [9]. Pichon N, Amiel JB, Franc¸ois B, Dugard A, Etchecopar C, Vignon P. Efficacy of and tolerance to mild induced hypothermia after out-of-hospital cardiac arrest using an endovascular cooling system. Crit Care 2007;11(3):R71. [10]. Tømte Ø, Drægni T, Mangschau A, Jacobsen D, Auestad B, Sunde K. A comparison of intravascular and surface cooling techniques in comatose cardiac arrest survivors. Crit Care Med 2011;39(3):443–9. [11]. Pittl U, Schratter A, Desch S, Diosteanu R, Lehmann D, Demmin K, et al. Invasive versus non-invasive cooling after in- and out-of-hospital cardiac arrest: a randomized trial. Clin Res Cardiol Off J Ger Card Soc 2013;102(8):607–14. [12]. Deye N, Cariou A, Girardie P, Pichon N, Megarbane B, Midez P, et al. Endovascular versus external targeted temperature management for patients with out-of-hospital cardiac arrest: a randomized, controlled study. Circulation 2015;132(3):182–93. [13]. Maze R, Le May MR, Froeschl M, Hazra SK, Wells PS, Osborne C, et al. Endovascular cooling catheter related thrombosis in patients undergoing therapeutic hypothermia for out of hospital cardiac arrest. Resuscitation 2014;85(10):1354–8. [14]. Müller A, Lorenz A, Seifert B, Keller E. Risk of thromboembolic events with endovascular cooling catheters in patients with subarachnoid hemorrhage. Neurocrit Care 2014;21(2):207–10. [15]. Mongardon N, Perbet S, Lemiale V, Dumas F, Poupet H, Charpentier J, et al. Infectious complications in out-of-hospital cardiac arrest patients in the therapeutic hypothermia era. Crit Care Med 2011;39(6):1359–64. [16]. Timsit JF, Farkas JC, Boyer JM, Martin JB, Misset B, Renaud B, et al. Central vein catheter-related thrombosis in intensive care patients: incidence, risks factors, and relationship with catheter-related sepsis. Chest 1998;114(1):207–13. [17]. Günther SC, Schwebel C, Hamidfar-Roy R, Bonadona A, Lugosi M, Ara-Somohano C, et al. Complications of intravascular catheters in ICU: definitions, incidence and severity: a randomized controlled trial comparing usual transparent dressings versus new-generation dressings (the ADVANCED study). Intensive Care Med 2016;42(11):1753–65. [18]. Parienti J-J, Mongardon N, Mégarbane B, Mira J-P, Kalfon P, Gros A, et al. Intravascular complications of central venous catheterization by insertion site. N Engl J Med 2015;373(13):1220–9. [19]. Minet C, Lugosi M, Savoye PY, Ruckly S, Bonadona A, Schwebel C, et al. Pulmonary emboism in mechanically ventilated patients requiring computed tomography: prevalence, risk factors, and outcome. Crit Care Med 2012;40(12):3202–8. [20]. Böttiger BW, Motsch J, Böhrer H, Böker T, Aulmann M, Nawroth PP, et al. Activation of blood coagulation after cardiac arrest is not balanced adequately by activation of endogenous fibrinolysis. Circulation 1995;92(9):2572–8. [21]. Adrie C, Monchi M, Laurent I, Um S, Yan SB, Thuong M, et al. Coagulopathy after successful cardiopulmonary resuscitation following cardiac arrest: implication of the protein C anticoagulant pathway. J Am Coll Cardiol 2005;46(1):21–8. [22]. MedDRA, https://www.meddra.org/. [23]. Fernandez A, Schmidt JM, Claassen J, Pavlicova M, Huddleston D, Kreiter KT, et al. Fever after subarachnoid hemorrhage: risk factors and impact on outcome. Neurology 2007;68(13):1013–9. [24]. Simosa HF, Petersen DJ, Agarwal SK, Burke PA, Hirsch EF. Increased risk of deep venous thrombosis with endovascular cooling in patients with traumatic head injury. Am Surg 2007;73(5):461–4. [25]. Reccius A, Mercado P, Vargas P, Canals C, Montes J. Inferior vena cava thrombosis related to hypothermia catheter: report of 20 consecutive cases. Neurocrit Care 2015;23(1):72–7. [26]. Jacob M, Hassager C, Bro-Jeppesen J, Ostrowski SR, Thomsen JH, Wanscher M, et al. The effect of targeted temperature management on coagulation parameters and bleeding events after out-of-hospital cardiac arrest of presumed cardiac cause. Resuscitation 2015;96:260–7. [27]. Nielsen AKW, Jeppesen AN, Kirkegaard H, Hvas A-M. Changes in coagulation during therapeutic hypothermia in cardiac arrest patients. Resuscitation 2016;98:85–90. [28]. Perbet S, Mongardon N, Dumas F, Bruel C, Lemiale V, Mourvillier B, et al. Earlyonset pneumonia after cardiac arrest: characteristics, risk factors and influence on prognosis. Am J Respir Crit Care Med 2011;184(9):1048–54. [29]. Kimura A, Sakurada S, Ohkuni H, Todome Y, Kurata K. Moderate hypothermia delays proinflammatory cytokine production of human peripheral blood mononuclear cells. Crit Care Med 2002;30(7):1499–502.

6

O. Andremont et al. / Resuscitation 124 (2018) 1–6

[30]. Dankiewicz J, Nielsen N, Linder A, Kuiper M, Wise MP, Cronberg T, et al. Infectious complications after out-of-hospital cardiac arrest-A comparison between two target temperatures. Resuscitation 2017;113:70–6. [31]. Rijnders BJA, Van Wijngaerden E, Peetermans WE. Catheter-tip colonization as a surrogate end point in clinical studies on catheter-related bloodstream infection: how strong is the evidence? Clin Infect Dis 2002;35(9):1053–8. [32]. Patel N, Nair SU, Gowd P, Gupta A, Morris D, Geronilla GG, et al. Central line associated blood stream infection related to cooling catheter in cardiac arrest survivors undergoing therapeutic hypothermia by endovascular cooling. Conn Med 2013;77(1):35–41. [33]. Shah N, Chaudhary R, Mehta K, Agarwal V, Garg J, Freudenberger R, et al. Therapeutic hypothermia and stent thrombosis: a nationwide analysis. JACC Cardiovasc Interv 2016;9(17):1801–11.

[34]. Gouffran G, Rosencher J, Bougouin W, Jakamy R, Joffre J, Lamhaut L, et al. Stent thrombosis after primary percutaneous coronary intervention in comatose survivors of out-of-hospital cardiac arrest: are the new P2Y12 inhibitors really more effective than clopidogrel? Resuscitation 2016;98:73–8. [35]. Cariou A, Deye N, Vivien B, Richard O, Pichon N, Bourg A, et al. Early highdose erythropoietin therapy after out-of-hospital cardiac arrest: a multicenter, randomized controlled trial. J Am Coll Cardiol 2016;68(1):40–9. [36]. Oh SH, Oh JS, Kim Y-M, Park KN, Choi SP, Kim GW, et al. An observational study of surface versus endovascular cooling techniques in cardiac arrest patients: a propensity-matched analysis. Crit Care 2015;19:85.