Resuscitation 91 (2015) 56–62
Contents lists available at ScienceDirect
Resuscitation journal homepage: www.elsevier.com/locate/resuscitation
Clinical Paper
Hemodynamic targets during therapeutic hypothermia after cardiac arrest: A prospective observational study夽 K. Ameloot a,∗,1 , I. Meex b,c,1 , C. Genbrugge b,c , F. Jans b,c , W. Boer b , D. Verhaert a , W. Mullens a,c , B. Ferdinande a , M. Dupont a , C. De Deyne b,c , J. Dens a,c a
Department of Cardiology, Ziekenhuis Oost-Limburg, Genk, Belgium Department of Anesthesiology and Critical Care Medicine, Ziekenhuis Oost-Limburg, Genk, Belgium c Faculty of Medicine and Life Sciences, University Hasselt, Diepenbeek, Belgium b
a r t i c l e
i n f o
Article history: Received 26 October 2014 Received in revised form 16 February 2015 Accepted 8 March 2015 Keywords: Cardiac arrest Hemodynamic targets Cerebrovascular circulation Therapeutic hypothermia
a b s t r a c t Aim: In analogy with sepsis, current post-cardiac arrest (CA) guidelines recommend to target mean arterial pressure (MAP) above 65 mmHg and SVO2 above 70%. This is unsupported by mortality or cerebral perfusion data. The aim of this study was to explore the associations between MAP, SVO2 , cerebral oxygenation and survival. Methods: Prospective, observational study during therapeutic hypothermia (24 h – 33 ◦ C) in 82 post-CA patients monitored with near-infrared spectroscopy. Results: Forty-three patients (52%) survived in CPC 1–2 until 180 days post-CA. The mean MAP range associated with maximal survival was 76–86 mmHg (OR 2.63, 95%CI [1.01; 6.88], p = 0.04). The mean SVO2 range associated with maximal survival was 67–72% (OR 8.23, 95%CI [2.07; 32.68], p = 0.001). In two separate multivariate models, a mean MAP (OR 3.72, 95% CI [1.11; 12.50], p = 0.03) and a mean SVO2 (OR 10.32, 95% CI [2.03; 52.60], p = 0.001) in the optimal range persisted as independently associated with increased survival. Based on more than 1 625 000 data points, we found a strong linear relation between SVO2 (range 40–90%) and average cerebral saturation (R2 0.86) and between MAP and average cerebral saturation for MAP’s between 45 and 101 mmHg (R2 0.83). Based on our hemodynamic model, the MAP and SVO2 ranges associated with optimal cerebral oxygenation were determined to be 87–101 mmHg and 70–75%. Conclusion: we showed that a MAP range between 76–86 mmHg and SVO2 range between 67% and 72% were associated with maximal survival. Optimal cerebral saturation was achieved with a MAP between 87–101 mmHg and a SVO2 between 70% and 75%. Prospective interventional studies are needed to investigate whether forcing MAP and SVO2 in the suggested range with additional pharmacological support would improve outcome. © 2015 Elsevier Ireland Ltd. All rights reserved.
1. Introduction The relationship between global hemodynamics, cerebral oxygenation and survival has not been investigated in post-cardiac arrest patients.1 In the absence of strong evidence, current guidelines are based on the assumption that the post-cardiac arrest syndrome is a sepsis-like syndrome. Therefore, it is recommended
夽 A Spanish translated version of the summary of this article appears as Appendix in the final online version at http://dx.doi.org/10.1016/j.resuscitation.2015.03.016. ∗ Corresponding author at: Ziekenhuis Oost-Limburg, Schiepse Bos, 3600 Genk, Belgium. Tel.: +32 89327106. E-mail address: [email protected] (K. Ameloot). 1 These authors contributed equally to this work. http://dx.doi.org/10.1016/j.resuscitation.2015.03.016 0300-9572/© 2015 Elsevier Ireland Ltd. All rights reserved.
to target mean arterial pressure (MAP) above 65 mmHg, systolic blood pressure above 90 mmHg and mixed venous oxygen saturation (SVO2 ) above 70%.2 However, the post-cardiac arrest syndrome is clearly a distinct and more complex entity than a sepsis-like syndrome alone, and aiming for the same hemodynamic goals is probably oversimplification. First, in critically ill patients without primary brain injury (e.g. sepsis), cerebral perfusion is kept stable in a broad range of blood pressures by cerebral autoregulation. In contrast, in a subset of post-cardiac arrest patients the lower threshold of cerebral autoregulation is shifted rightward and these patients might benefit from resuscitation to higher MAP’s.3 Second, patients with a reduced left ventricular function might benefit from afterload reduction to maintain stroke volume and cerebral perfusion.4 The optimal MAP should maintain cerebral perfusion without exposing the damaged myocardium to excessive
K. Ameloot et al. / Resuscitation 91 (2015) 56–62
afterload.5 Therefore, the aims of this prospective observational study were to explore the relationships between global hemodynamics, cerebral oxygenation and outcome in post-cardiac arrest patients during therapeutic hypothermia in the first 24 h after ICU admission. 2. Methods 2.1. Study population All comatose survivors after non-traumatic cardiac arrest treated in our tertiary care hospital (Ziekenhuis Oost-Limburg, Genk, Belgium) were prospectively enrolled in this study. Patients could have been resuscitated in-hospital, referred by another hospital or admitted by our emergency ward. All patients were treated uniformly according to the institutional post-cardiac arrest protocol. Patients were routinely monitored by cerebral saturation and invasive arterial blood pressure monitoring on hospital arrival and by pulmonary artery catheter (PAC) shortly after admission on the cardiac intensive care unit unless PAC placement was contra-indicated or considered inappropriate by treating physicians. Written informed consent was obtained from a next of kin. Five patients with refractory shock who died during the first 24 h were excluded from analysis. The study protocol was approved by the local medical ethics committee. 2.2. General management Our post-cardiac arrest protocol has been described previously.7 Shortly, all patients were intubated, mechanically ventilated and sedated with propofol and remifentanil if hemodynamically tolerated. Cisatracurium was administrated in case of shivering during hypothermia. Unless an obvious non-cardiac cause could be identified, all patients were referred for urgent coronary angiography. Therapeutic hypothermia was induced in all patients shortly after admission by cold saline (4 ◦ C – 30 ml/kg) and further mechanically maintained in the ICU by endovascular (Icy-catheter, CoolGard® 3000, Alsius, Irvine, CA, USA) or surface (ArcticGelTM pads, Arctic Sun® 5000, Medivance, Louisville, Colorado, USA) cooling systems at 33 ◦ C for 24-h. After rewarming (0.3 ◦ C/h) sedation was titrated toward patient’s comfort. Patients were extubated after sufficient recovery. 2.3. Cerebral saturation monitoring Cerebral tissue oxygen saturation was continuously measured with near infrared spectroscopy (NIRS), using the FORE-SIGHTTM technology (CAS Medical Systems, Branford, CT, USA). Sensors were bilaterally applied to each frontotemporal area before the start of mechanically induced hypothermia. Sensors were covered to prevent ambient light interference. Cerebral saturation data were transmitted electronically to a personal computer with a 2 s time interval. Cerebral saturation data were not used to guide any form of hemodynamic management. 2.4. Hemodynamic monitoring and management Patients were treated according to the recommended guidelines.2 If signs of inadequate circulation persisted despite correct fluid resuscitation (wedge pressure > 18 mmHg), norepinephrine was infused with a target MAP > 65 mmHg and subsequently dobutamine to target a cardiac index > 2.2 L/min/m2 . An IABP was installed as deemed necessary by treating physicians. A transthoracic echocardiography was performed in the first 24 h after admission. Continuous thermodilution cardiac output and
57
SVO2 were measured continuously by a new generation PAC (CCOmbo PAC® , Edwards Life Science, Irvine, CA, USA) connected to the appropriate monitor (Vigilance II® , Edwards Life Science, Irvine, CA, USA). The continuous SVO2 monitoring system was calibrated as prescribed by the company. These data were transmitted electronically to a personal computer with a 2 s time interval, together with information on blood temperature, oxygen saturation, MAP obtained from a radial artery line and cerebral saturation data. In the first 29 patients enrolled in the database, the hemodynamic data were written down manually every 15 min. 2.5. Statistics Results are expressed as mean (±SD, standard deviation) unless otherwise stated. Survival was defined as survival in CPC 1–2 until 180 days post-CA. First, for each patient, the mean SVO2 and MAP were calculated by averaging all obtained values during the first 24 h after the first SVO2 value (or MAP value in patients not monitored with a PAC). To determine the MAP and SVO2 associated with maximal survival, odds ratios (and 95% confidence intervals) were calculated per 5% SVO2 and per 10 mmHg MAP intervals. To test for significance, a Chi-square test was performed for the interval with the highest odd’s ratio (all expected frequencies were more than 5). Additionally, the percentage of time in the obtained optimal SVO2 and MAP range was calculated for each patient. Univariate logistic regression was used to test for significance (assuming that a higher percentage of time spent in the suggested SVO2 or MAP range would result in better survival). Similarly, other candidate binary variables were evaluated by a Chi-square test and univariate logistic regression was used for continuous candidate variables. A multivariate model was constructed using backward multivariate logistic regression with all candidate variables (age, bystander CPR < 10 min, shockable rhythm, initial lactate, mean lactate, left ventricular ejection fraction). A separate multivariate model was constructed for MAP and SVO2 . Second, patients were stratified to be in the low/optimal/high MAP and SVO2 subgroups according to their average values during the first 24 h after admission. These subgroups were compared using one-way ANOVA. Third, to construct a hemodynamic model, the average cerebral saturation was calculated per mmHg MAP (range 45–110 mmHg) and per % SVO2 (range 40–90%) and the average SVO2 was calculated per mmHg MAP. Univariate linear regression with calculation of Pearson’s correlation coefficient was used to describe the relationships between MAP, SVO2 and cerebral saturation. Assuming that the optimal MAP range based on this hemodynamic model has to maintain cerebral oxygenation without exposing the damaged myocardium to excessive afterload, the lowest optimal MAP was defined as the lowest MAP with a corresponding cerebral saturation equal to the average of all cerebral saturations associated with all higher MAP’s than the taken MAP value. The highest optimal MAP was defined as the MAP associated with maximal cerebral saturation. The optimal hemodynamic SVO2 range was considered to be corresponding with this optimal MAP range. Statistical analysis was performed using Matlab software (version R2010b, Mathworks, USA). A p-value 65 mmHg and SVO2 > 70%.2 In this prospective, observational
60
K. Ameloot et al. / Resuscitation 91 (2015) 56–62
Fig. 2. Relationship between mixed venous oxygen saturation (SVO2 ) and cerebral saturation.
Fig. 3. Relationship between mean arterial pressure (MAP), mixed venous oxygen saturation (SVO2 ) and cerebral saturation.
study in 82 post-CA patients during therapeutic hypothermia (33 ◦ C) in the first 24 h after ICU admission, we showed that a MAP range of 76–86 mmHg and SVO2 range of 67–72% were associated with maximal survival. Based on our hemodynamic model, the optimal cerebral saturation was determined to be 67–68%. This was achieved with a MAP between 87–101 mmHg and a SVO2 between 70% and 75%. The recommended MAP of 65 mmHg did not result in this optimal cerebral saturation value.
First, we defined the optimal MAP (for the first 24 h post-CA and during TH (33 ◦ C)) associated with maximal survival to be between 76 and 86 mmHg. We showed that patients with a mean MAP in this range had a 2.6-fold increased survival chance. Moreover, we found that the probability of survival increased with 4% per extra percentage of time in the optimal MAP range. In contrast, relying on our hemodynamic model, the currently recommended MAP of 65 mmHg would result in a cerebral saturation of 65% which is 2% below our suggested threshold for autoregulation. The clinical relevance of this small cerebral desaturation is illustrated by our observation that the achievement of a MAP above this arbitrary cut-off did not increase the probability of survival. Therefore, these one-size-fits-all hemodynamic targets for all critical ill patients may lead to under-resuscitation, cerebral desaturation and increased mortality. Importantly, patients who died during the first 24 h due to refractory post-resuscitation shock were excluded in our study because we think post-resuscitation targets should be designed for patients with the largest potential benefit of goal directed hemodynamic optimization (i.e. potential survivors at risk for additional cerebral damage due to hypo- or hyperperfusion). Our findings are in agreement with 2 retrospective and 1 prospective study previously published on the positive association between higher MAP’s during the first 6 h post-ROSC and increased survival in post-cardiac arrest patients.11–13 Additionally, the study by Kilgannon et al. also failed to show a further improvement of the outcome in patients with MAP’s above 90 mmHg. This is in agreement with our data and may be related to a decrease in cerebral oxygenation at MAPs above 101 mmHg. On the other hand, a sub-study of the target temperature management trial showed that mortality increases in patients with an average MAP below 65 mmHg.19 Second, we showed that patients with a mean SVO2 between 67% and 72% during the first 24 h after admission had an eightfold increased survival chance. We also found that the probability of survival increased with 3% per extra percentage of time in this optimal SVO2 range. This suggests that every unit of time out of the suggested target range may contribute to cumulative cerebral damage. Indeed, patients with a mean SVO2 above or below the target range significantly died more often due to neurological failure. Survival was poor in the subgroup of patients with a suboptimal cerebral saturation as a result of insufficient cardiac output. Additionally, we observed the highest mortality rates in the subgroup of patients with the highest SVO2 and the highest cerebral saturation. The association between a hyperdynamic circulation and increased mortality has not been described previously in postcardiac arrest patients. This surprising finding cannot be explained by arterial hyperoxygenation since arterial pO2 levels were equal in subgroups with low, normal and high SVO2 ’s. Three potential explanations apply. First, a hyperdynamic circulation may cause cerebral hypersaturation which may cause further brain damage due to the toxic effects of highly reactive free oxygen radicals. This is supported by previous studies on the detrimental effect of arterial hyperoxemia post-cardiac arrest.8,9 Second, fewer patients in this high-SVO2 subgroup presented with a shockable rhythm and received early bystander CPR. Therefore, the longer no-flow time in this patient group may have induced a more severe systemic inflammatory response syndrome (SIRS). In this case, there would be no causality and a high SVO2 would be a marker of brain damage. However, in our multivariate model, a mean SVO2 in the optimal range persisted as a predictor of outcome after correction for bystander CPR and shockable rhythm. Finally, high SVO2 values may be an expression of reduced oxygen consumption. This may again be due to severe brain damage (‘functional brain dead’).7 Third, we showed that the cerebral saturation is extremely dependent on cardiac output for which we take SVO2 as a surrogate measure. Cardiac output by itself was not studied since we
K. Ameloot et al. / Resuscitation 91 (2015) 56–62
recently showed large inaccuracy regarding thermodilution cardiac output measurements during therapeutic hypothermia.10 This linear relationship between SVO2 and cerebral saturation was previously described during cardiac surgery.17 In the present study, cerebral saturation was also strongly dependent on MAP, although we found MAP’s above 101 mmHg were associated with progressively decreasing cerebral saturations. Two potential mechanisms apply for this plateau-phase. First, it seems likely that the increased left ventricular afterload at higher MAP’s mechanically impedes a further increase in cardiac output or that the increased blood pressure activates cardio-depressive feedback loops.6 This is supported by our observation that also SVO2 progressively decreases at MAP’s higher than 101 mmHg. Second, it may also be that cerebral autoregulation is indeed shifted rightward as suggested in a previous study with transcranial Doppler in normothermic post-cardiac arrest patients.3 In this study, the lower limit of autoregulation was suggested to be between 80 and 120 mmHg. In contrast, Bouzat et al. used norepinephrine to increase MAP from 72 to 90 mmHg in 10 post-cardiac arrest patients during the maintenance phase of therapeutic hypothermia.7 They did not observe a significant increase in cerebral saturation (from 73.6% to 74.1%) and concluded that cerebrovascular autoregulation was preserved in hypothermic patients in the studied MAP range. This is the first study to focus on the relationship between global hemodynamics, cerebral oxygenation and clinical outcome in postCA patients. We acknowledge that our study has some important limitations to consider. This is a small single center observational trial and our results may be underpowered. On the other hand, our hemodynamic model is based on more than 1.6 million data points. Additionally, a post hoc power calculation showed that with a given probability of good outcome of 83% in the optimal SVO2 subgroup and of 37% in patients with lower and higher average SVO2 ’s, our study sample would enable to detect at least a 10% survival difference with a power of 94% at a significance level of 0.05. A second limitation is that this was not an interventional trial comparing different MAP and SVO2 targets. It remains unknown whether forcing MAP or SVO2 into the suggested range with additional pharmacological support would improve outcome. However, our results do set the scene for a future interventional trial targeting SVO2 and MAP in the suggested range. In a previous study by Gaieski et al. 20 prospective post-cardiac arrest patients received early goal directed hemodynamic optimization with a target central venous saturation above 65% and MAP between 80 and 100 mmHg.14 A trend toward better survival was observed in treated patients (50% survival) when compared with 18 historic controls (22% survival). A third limitation is that we do not have detailed information about vasopressor use during the study period. However, in principle, all patients were treated by the same dedicated staff instructed to maintain MAP above 65 mmHg with the lowest effective dose of pharmacological support. Fluid status and hemodynamic support were optimized every 30 min. This makes it very unlikely that mean MAP’s in the suggested target range over the 24-h study period were due to a systematic overuse of vasopressors. Fourth, we acknowledge that our strategy to perform serial Chisquare testing to determine the optimal MAP and SVO2 intervals should be regarded as a form of exploratory statistics that should be confirmed in a second validation cohort. Fifth, we used NIRS as a surrogate measure for cerebral perfusion since this is easy to use, non-invasive and allows beat-to-beat assessment of endorgan perfusion. Some authors prefer to use transcranial Doppler as a gold standard to assess cerebral perfusion.3 NIRS was previously validated against CT-perfusion for the assessment of regional cerebral perfusion in patients with traumatic brain injury.18 Sixth, we included patients admitted from in and out-of hospital CA of which some were referred by nearby hospitals and of which some spent a period in the catherization lab. Since we could only start
61
hemodynamic recordings after ICU admission and PAC placement, there may be small heterogeneity in the time interval between ROSC and start of the study. However, we do not believe that this influenced our results since we studied 24 h hemodynamics and the maximal time delay between ROSC and start of the study is estimated to be around 3 h. Finally, this study was conducted in the hypothermia era. A recent trial failed to show benefit of hypothermia on top of targeted temperature management.15 It’s uncertain whether our suggested hemodynamic targets would still hold in normothermic patients. It has been suggested that the MAP associated with optimal cerebral perfusion may be altered by therapeutic hypothermia.16 In conclusion, we showed that a MAP range between 76–86 mmHg and SVO2 range between 67% and 72% were associated with maximal survival. Based on our hemodynamic model, the optimal cerebral saturation was determined to be 67–68%. This was achieved with a MAP between 87–101 mmHg and a SVO2 between 70% and 75%. Prospective interventional studies are needed to investigate whether forcing MAP and SVO2 in the suggested range with additional pharmacological support would improve outcome. Conflict of interest statement The authors do not have anything to disclose. Acknowledgements The authors would like to thank Annelies Gerits and An Creemers, our consulting statisticians. References 1. Jones AE, Shapiro NI, Kilgannon JH, Trzeciak S, Emergency Medicine Shock Research Network (EMSHOCKNET) investigators. Goal-directed hemodynamic optimization in the post-cardiac arrest syndrome: a systematic review. Resuscitation 2008;77:26–9. 2. Peberdy MA, Callaway CW, Neumar RW, et al. Post-cardiac arrest care: 2010 American Heart Association guidelines for cardiopulmonary resuscitation and emergency cardiovascular care. Circulation 2010;122:S768–86. 3. Sundgreen C, Larsen FS, Herzog TM, Knudsen GM, Boesgaard S, Aldershvile J. Autoregulation of cerebral blood flow in patients resuscitated from cardiac arrest. Stroke 2001;32:128–32. 4. Badeer HS. Contractile tension in the myocardium. Am Heart J 1963;66:432–4. 5. Nolan JP, Soar J. Post resuscitation care – time for a bundle? Resuscitation 2008;76:161–2. 6. Meex I, Dens J, Jans F, et al. Cerebral tissue oxygen saturation during therapeutic hypothermia in post-cardiac arrest patients. Resuscitation 2013;84:788–93. 7. Bouzat P, Suys T, Sala N, Oddo M. Effect of moderate hyperventilation and induced hypertension on cerebral tissue oxygenation after cardiac arrest and therapeutic hypothermia. Resuscitation 2013;84:1540–5. 8. Kilgannon JH, Jones AE, Parrillo JE, et al. Relationship between supranormal oxygen tension and outcome after resuscitation from cardiac arrest. Circulation 2011;123:2717–22. 9. Kilgannon JH, Jones AE, Shapiro NI, et al. Association between arterial hyperoxia following resuscitation from cardiac arrest and in-hospital mortality. JAMA 2010;303:2165–71. 10. Ameloot K, Meex I, Genbrugge C, et al. Accuracy of continuous thermodilution cardiac output monitoring by pulmonary artery catheter during therapeutic hypothermia in post-cardiac arrest patients. Resuscitation 2014;85:1263–8. 11. Beylin ME, Perman SM, Abella BS, et al. Higher mean arterial pressure with or without vasoactive agents is associated with increased survival and better neurological outcomes in comatose survivors of cardiac arrest. Intensive Care Med 2013;39:1981–8. 12. Trzeciak S, Jones AE, Kilgannon JH, et al. Significance of arterial hypotension after resuscitation from cardiac arrest. Crit Care Med 2009;37:2895–903. 13. Kilgannon JH, Roberts BW, Jones AE, et al. Arterial blood pressure and neurological outcome after resuscitation from cardiac arrest. Crit Care Med 2014;42:2083–91. 14. Gaieski DF, Band RA, Abella BS, et al. Early goal-directed hemodynamic optimization combined with therapeutic hypothermia in comatose survivors of out-of-hospital cardiac arrest. Early goal-directed hemodynamic optimization combined with therapeutic hypothermia in comatose survivors of out-ofhospital cardiac arrest. Resuscitation 2009;80:418–24. 15. Nielsen N, Wetterslev J, Cronberg T, et al. Targeted temperature management at 33◦ C versus 36◦ C after cardiac arrest. N Engl J Med 2013;369:2197–206.
62
K. Ameloot et al. / Resuscitation 91 (2015) 56–62
16. Lee JK, Brady KM, Mytar JO, et al. Cerebral blood flow and cerebrovascular autoregulation in a swine model of pediatric cardiac arrest and hypothermia. Crit Care Med 2011;39:2337–45. 17. Paarmann H, Heringlake M, Heinze H, et al. Non-invasive cerebral oxygenation reflects mixed venous oxygen saturation during the varying haemodynamic conditions in patients undergoing transapical transcatheter aortic valve implantation. Interact Cardiovasc Thorac Surg 2012;14:268–72.
18. Taussky P, O’Neal B, Daugherty WP, et al. Validation of frontal near-infrared spectroscopy as noninvasive bedside monitoring for regional cerebral blood flow in brain-injured patients. Neurosurg Focus 2012;32:E2. 19. Bro-Jeppesen J, Annborn M, Hassager C, et al. Hemodynamics and vasopressor support during targeted temperature management at 33◦ C versus 36◦ C after out-of-hospital cardiac arrest: a post hoc study of the target temperature management trial. Crit Care Med 2015;43:318–27.
Contents lists available at ScienceDirect
Resuscitation journal homepage: www.elsevier.com/locate/resuscitation
Clinical Paper
Hemodynamic targets during therapeutic hypothermia after cardiac arrest: A prospective observational study夽 K. Ameloot a,∗,1 , I. Meex b,c,1 , C. Genbrugge b,c , F. Jans b,c , W. Boer b , D. Verhaert a , W. Mullens a,c , B. Ferdinande a , M. Dupont a , C. De Deyne b,c , J. Dens a,c a
Department of Cardiology, Ziekenhuis Oost-Limburg, Genk, Belgium Department of Anesthesiology and Critical Care Medicine, Ziekenhuis Oost-Limburg, Genk, Belgium c Faculty of Medicine and Life Sciences, University Hasselt, Diepenbeek, Belgium b
a r t i c l e
i n f o
Article history: Received 26 October 2014 Received in revised form 16 February 2015 Accepted 8 March 2015 Keywords: Cardiac arrest Hemodynamic targets Cerebrovascular circulation Therapeutic hypothermia
a b s t r a c t Aim: In analogy with sepsis, current post-cardiac arrest (CA) guidelines recommend to target mean arterial pressure (MAP) above 65 mmHg and SVO2 above 70%. This is unsupported by mortality or cerebral perfusion data. The aim of this study was to explore the associations between MAP, SVO2 , cerebral oxygenation and survival. Methods: Prospective, observational study during therapeutic hypothermia (24 h – 33 ◦ C) in 82 post-CA patients monitored with near-infrared spectroscopy. Results: Forty-three patients (52%) survived in CPC 1–2 until 180 days post-CA. The mean MAP range associated with maximal survival was 76–86 mmHg (OR 2.63, 95%CI [1.01; 6.88], p = 0.04). The mean SVO2 range associated with maximal survival was 67–72% (OR 8.23, 95%CI [2.07; 32.68], p = 0.001). In two separate multivariate models, a mean MAP (OR 3.72, 95% CI [1.11; 12.50], p = 0.03) and a mean SVO2 (OR 10.32, 95% CI [2.03; 52.60], p = 0.001) in the optimal range persisted as independently associated with increased survival. Based on more than 1 625 000 data points, we found a strong linear relation between SVO2 (range 40–90%) and average cerebral saturation (R2 0.86) and between MAP and average cerebral saturation for MAP’s between 45 and 101 mmHg (R2 0.83). Based on our hemodynamic model, the MAP and SVO2 ranges associated with optimal cerebral oxygenation were determined to be 87–101 mmHg and 70–75%. Conclusion: we showed that a MAP range between 76–86 mmHg and SVO2 range between 67% and 72% were associated with maximal survival. Optimal cerebral saturation was achieved with a MAP between 87–101 mmHg and a SVO2 between 70% and 75%. Prospective interventional studies are needed to investigate whether forcing MAP and SVO2 in the suggested range with additional pharmacological support would improve outcome. © 2015 Elsevier Ireland Ltd. All rights reserved.
1. Introduction The relationship between global hemodynamics, cerebral oxygenation and survival has not been investigated in post-cardiac arrest patients.1 In the absence of strong evidence, current guidelines are based on the assumption that the post-cardiac arrest syndrome is a sepsis-like syndrome. Therefore, it is recommended
夽 A Spanish translated version of the summary of this article appears as Appendix in the final online version at http://dx.doi.org/10.1016/j.resuscitation.2015.03.016. ∗ Corresponding author at: Ziekenhuis Oost-Limburg, Schiepse Bos, 3600 Genk, Belgium. Tel.: +32 89327106. E-mail address: [email protected] (K. Ameloot). 1 These authors contributed equally to this work. http://dx.doi.org/10.1016/j.resuscitation.2015.03.016 0300-9572/© 2015 Elsevier Ireland Ltd. All rights reserved.
to target mean arterial pressure (MAP) above 65 mmHg, systolic blood pressure above 90 mmHg and mixed venous oxygen saturation (SVO2 ) above 70%.2 However, the post-cardiac arrest syndrome is clearly a distinct and more complex entity than a sepsis-like syndrome alone, and aiming for the same hemodynamic goals is probably oversimplification. First, in critically ill patients without primary brain injury (e.g. sepsis), cerebral perfusion is kept stable in a broad range of blood pressures by cerebral autoregulation. In contrast, in a subset of post-cardiac arrest patients the lower threshold of cerebral autoregulation is shifted rightward and these patients might benefit from resuscitation to higher MAP’s.3 Second, patients with a reduced left ventricular function might benefit from afterload reduction to maintain stroke volume and cerebral perfusion.4 The optimal MAP should maintain cerebral perfusion without exposing the damaged myocardium to excessive
K. Ameloot et al. / Resuscitation 91 (2015) 56–62
afterload.5 Therefore, the aims of this prospective observational study were to explore the relationships between global hemodynamics, cerebral oxygenation and outcome in post-cardiac arrest patients during therapeutic hypothermia in the first 24 h after ICU admission. 2. Methods 2.1. Study population All comatose survivors after non-traumatic cardiac arrest treated in our tertiary care hospital (Ziekenhuis Oost-Limburg, Genk, Belgium) were prospectively enrolled in this study. Patients could have been resuscitated in-hospital, referred by another hospital or admitted by our emergency ward. All patients were treated uniformly according to the institutional post-cardiac arrest protocol. Patients were routinely monitored by cerebral saturation and invasive arterial blood pressure monitoring on hospital arrival and by pulmonary artery catheter (PAC) shortly after admission on the cardiac intensive care unit unless PAC placement was contra-indicated or considered inappropriate by treating physicians. Written informed consent was obtained from a next of kin. Five patients with refractory shock who died during the first 24 h were excluded from analysis. The study protocol was approved by the local medical ethics committee. 2.2. General management Our post-cardiac arrest protocol has been described previously.7 Shortly, all patients were intubated, mechanically ventilated and sedated with propofol and remifentanil if hemodynamically tolerated. Cisatracurium was administrated in case of shivering during hypothermia. Unless an obvious non-cardiac cause could be identified, all patients were referred for urgent coronary angiography. Therapeutic hypothermia was induced in all patients shortly after admission by cold saline (4 ◦ C – 30 ml/kg) and further mechanically maintained in the ICU by endovascular (Icy-catheter, CoolGard® 3000, Alsius, Irvine, CA, USA) or surface (ArcticGelTM pads, Arctic Sun® 5000, Medivance, Louisville, Colorado, USA) cooling systems at 33 ◦ C for 24-h. After rewarming (0.3 ◦ C/h) sedation was titrated toward patient’s comfort. Patients were extubated after sufficient recovery. 2.3. Cerebral saturation monitoring Cerebral tissue oxygen saturation was continuously measured with near infrared spectroscopy (NIRS), using the FORE-SIGHTTM technology (CAS Medical Systems, Branford, CT, USA). Sensors were bilaterally applied to each frontotemporal area before the start of mechanically induced hypothermia. Sensors were covered to prevent ambient light interference. Cerebral saturation data were transmitted electronically to a personal computer with a 2 s time interval. Cerebral saturation data were not used to guide any form of hemodynamic management. 2.4. Hemodynamic monitoring and management Patients were treated according to the recommended guidelines.2 If signs of inadequate circulation persisted despite correct fluid resuscitation (wedge pressure > 18 mmHg), norepinephrine was infused with a target MAP > 65 mmHg and subsequently dobutamine to target a cardiac index > 2.2 L/min/m2 . An IABP was installed as deemed necessary by treating physicians. A transthoracic echocardiography was performed in the first 24 h after admission. Continuous thermodilution cardiac output and
57
SVO2 were measured continuously by a new generation PAC (CCOmbo PAC® , Edwards Life Science, Irvine, CA, USA) connected to the appropriate monitor (Vigilance II® , Edwards Life Science, Irvine, CA, USA). The continuous SVO2 monitoring system was calibrated as prescribed by the company. These data were transmitted electronically to a personal computer with a 2 s time interval, together with information on blood temperature, oxygen saturation, MAP obtained from a radial artery line and cerebral saturation data. In the first 29 patients enrolled in the database, the hemodynamic data were written down manually every 15 min. 2.5. Statistics Results are expressed as mean (±SD, standard deviation) unless otherwise stated. Survival was defined as survival in CPC 1–2 until 180 days post-CA. First, for each patient, the mean SVO2 and MAP were calculated by averaging all obtained values during the first 24 h after the first SVO2 value (or MAP value in patients not monitored with a PAC). To determine the MAP and SVO2 associated with maximal survival, odds ratios (and 95% confidence intervals) were calculated per 5% SVO2 and per 10 mmHg MAP intervals. To test for significance, a Chi-square test was performed for the interval with the highest odd’s ratio (all expected frequencies were more than 5). Additionally, the percentage of time in the obtained optimal SVO2 and MAP range was calculated for each patient. Univariate logistic regression was used to test for significance (assuming that a higher percentage of time spent in the suggested SVO2 or MAP range would result in better survival). Similarly, other candidate binary variables were evaluated by a Chi-square test and univariate logistic regression was used for continuous candidate variables. A multivariate model was constructed using backward multivariate logistic regression with all candidate variables (age, bystander CPR < 10 min, shockable rhythm, initial lactate, mean lactate, left ventricular ejection fraction). A separate multivariate model was constructed for MAP and SVO2 . Second, patients were stratified to be in the low/optimal/high MAP and SVO2 subgroups according to their average values during the first 24 h after admission. These subgroups were compared using one-way ANOVA. Third, to construct a hemodynamic model, the average cerebral saturation was calculated per mmHg MAP (range 45–110 mmHg) and per % SVO2 (range 40–90%) and the average SVO2 was calculated per mmHg MAP. Univariate linear regression with calculation of Pearson’s correlation coefficient was used to describe the relationships between MAP, SVO2 and cerebral saturation. Assuming that the optimal MAP range based on this hemodynamic model has to maintain cerebral oxygenation without exposing the damaged myocardium to excessive afterload, the lowest optimal MAP was defined as the lowest MAP with a corresponding cerebral saturation equal to the average of all cerebral saturations associated with all higher MAP’s than the taken MAP value. The highest optimal MAP was defined as the MAP associated with maximal cerebral saturation. The optimal hemodynamic SVO2 range was considered to be corresponding with this optimal MAP range. Statistical analysis was performed using Matlab software (version R2010b, Mathworks, USA). A p-value 65 mmHg and SVO2 > 70%.2 In this prospective, observational
60
K. Ameloot et al. / Resuscitation 91 (2015) 56–62
Fig. 2. Relationship between mixed venous oxygen saturation (SVO2 ) and cerebral saturation.
Fig. 3. Relationship between mean arterial pressure (MAP), mixed venous oxygen saturation (SVO2 ) and cerebral saturation.
study in 82 post-CA patients during therapeutic hypothermia (33 ◦ C) in the first 24 h after ICU admission, we showed that a MAP range of 76–86 mmHg and SVO2 range of 67–72% were associated with maximal survival. Based on our hemodynamic model, the optimal cerebral saturation was determined to be 67–68%. This was achieved with a MAP between 87–101 mmHg and a SVO2 between 70% and 75%. The recommended MAP of 65 mmHg did not result in this optimal cerebral saturation value.
First, we defined the optimal MAP (for the first 24 h post-CA and during TH (33 ◦ C)) associated with maximal survival to be between 76 and 86 mmHg. We showed that patients with a mean MAP in this range had a 2.6-fold increased survival chance. Moreover, we found that the probability of survival increased with 4% per extra percentage of time in the optimal MAP range. In contrast, relying on our hemodynamic model, the currently recommended MAP of 65 mmHg would result in a cerebral saturation of 65% which is 2% below our suggested threshold for autoregulation. The clinical relevance of this small cerebral desaturation is illustrated by our observation that the achievement of a MAP above this arbitrary cut-off did not increase the probability of survival. Therefore, these one-size-fits-all hemodynamic targets for all critical ill patients may lead to under-resuscitation, cerebral desaturation and increased mortality. Importantly, patients who died during the first 24 h due to refractory post-resuscitation shock were excluded in our study because we think post-resuscitation targets should be designed for patients with the largest potential benefit of goal directed hemodynamic optimization (i.e. potential survivors at risk for additional cerebral damage due to hypo- or hyperperfusion). Our findings are in agreement with 2 retrospective and 1 prospective study previously published on the positive association between higher MAP’s during the first 6 h post-ROSC and increased survival in post-cardiac arrest patients.11–13 Additionally, the study by Kilgannon et al. also failed to show a further improvement of the outcome in patients with MAP’s above 90 mmHg. This is in agreement with our data and may be related to a decrease in cerebral oxygenation at MAPs above 101 mmHg. On the other hand, a sub-study of the target temperature management trial showed that mortality increases in patients with an average MAP below 65 mmHg.19 Second, we showed that patients with a mean SVO2 between 67% and 72% during the first 24 h after admission had an eightfold increased survival chance. We also found that the probability of survival increased with 3% per extra percentage of time in this optimal SVO2 range. This suggests that every unit of time out of the suggested target range may contribute to cumulative cerebral damage. Indeed, patients with a mean SVO2 above or below the target range significantly died more often due to neurological failure. Survival was poor in the subgroup of patients with a suboptimal cerebral saturation as a result of insufficient cardiac output. Additionally, we observed the highest mortality rates in the subgroup of patients with the highest SVO2 and the highest cerebral saturation. The association between a hyperdynamic circulation and increased mortality has not been described previously in postcardiac arrest patients. This surprising finding cannot be explained by arterial hyperoxygenation since arterial pO2 levels were equal in subgroups with low, normal and high SVO2 ’s. Three potential explanations apply. First, a hyperdynamic circulation may cause cerebral hypersaturation which may cause further brain damage due to the toxic effects of highly reactive free oxygen radicals. This is supported by previous studies on the detrimental effect of arterial hyperoxemia post-cardiac arrest.8,9 Second, fewer patients in this high-SVO2 subgroup presented with a shockable rhythm and received early bystander CPR. Therefore, the longer no-flow time in this patient group may have induced a more severe systemic inflammatory response syndrome (SIRS). In this case, there would be no causality and a high SVO2 would be a marker of brain damage. However, in our multivariate model, a mean SVO2 in the optimal range persisted as a predictor of outcome after correction for bystander CPR and shockable rhythm. Finally, high SVO2 values may be an expression of reduced oxygen consumption. This may again be due to severe brain damage (‘functional brain dead’).7 Third, we showed that the cerebral saturation is extremely dependent on cardiac output for which we take SVO2 as a surrogate measure. Cardiac output by itself was not studied since we
K. Ameloot et al. / Resuscitation 91 (2015) 56–62
recently showed large inaccuracy regarding thermodilution cardiac output measurements during therapeutic hypothermia.10 This linear relationship between SVO2 and cerebral saturation was previously described during cardiac surgery.17 In the present study, cerebral saturation was also strongly dependent on MAP, although we found MAP’s above 101 mmHg were associated with progressively decreasing cerebral saturations. Two potential mechanisms apply for this plateau-phase. First, it seems likely that the increased left ventricular afterload at higher MAP’s mechanically impedes a further increase in cardiac output or that the increased blood pressure activates cardio-depressive feedback loops.6 This is supported by our observation that also SVO2 progressively decreases at MAP’s higher than 101 mmHg. Second, it may also be that cerebral autoregulation is indeed shifted rightward as suggested in a previous study with transcranial Doppler in normothermic post-cardiac arrest patients.3 In this study, the lower limit of autoregulation was suggested to be between 80 and 120 mmHg. In contrast, Bouzat et al. used norepinephrine to increase MAP from 72 to 90 mmHg in 10 post-cardiac arrest patients during the maintenance phase of therapeutic hypothermia.7 They did not observe a significant increase in cerebral saturation (from 73.6% to 74.1%) and concluded that cerebrovascular autoregulation was preserved in hypothermic patients in the studied MAP range. This is the first study to focus on the relationship between global hemodynamics, cerebral oxygenation and clinical outcome in postCA patients. We acknowledge that our study has some important limitations to consider. This is a small single center observational trial and our results may be underpowered. On the other hand, our hemodynamic model is based on more than 1.6 million data points. Additionally, a post hoc power calculation showed that with a given probability of good outcome of 83% in the optimal SVO2 subgroup and of 37% in patients with lower and higher average SVO2 ’s, our study sample would enable to detect at least a 10% survival difference with a power of 94% at a significance level of 0.05. A second limitation is that this was not an interventional trial comparing different MAP and SVO2 targets. It remains unknown whether forcing MAP or SVO2 into the suggested range with additional pharmacological support would improve outcome. However, our results do set the scene for a future interventional trial targeting SVO2 and MAP in the suggested range. In a previous study by Gaieski et al. 20 prospective post-cardiac arrest patients received early goal directed hemodynamic optimization with a target central venous saturation above 65% and MAP between 80 and 100 mmHg.14 A trend toward better survival was observed in treated patients (50% survival) when compared with 18 historic controls (22% survival). A third limitation is that we do not have detailed information about vasopressor use during the study period. However, in principle, all patients were treated by the same dedicated staff instructed to maintain MAP above 65 mmHg with the lowest effective dose of pharmacological support. Fluid status and hemodynamic support were optimized every 30 min. This makes it very unlikely that mean MAP’s in the suggested target range over the 24-h study period were due to a systematic overuse of vasopressors. Fourth, we acknowledge that our strategy to perform serial Chisquare testing to determine the optimal MAP and SVO2 intervals should be regarded as a form of exploratory statistics that should be confirmed in a second validation cohort. Fifth, we used NIRS as a surrogate measure for cerebral perfusion since this is easy to use, non-invasive and allows beat-to-beat assessment of endorgan perfusion. Some authors prefer to use transcranial Doppler as a gold standard to assess cerebral perfusion.3 NIRS was previously validated against CT-perfusion for the assessment of regional cerebral perfusion in patients with traumatic brain injury.18 Sixth, we included patients admitted from in and out-of hospital CA of which some were referred by nearby hospitals and of which some spent a period in the catherization lab. Since we could only start
61
hemodynamic recordings after ICU admission and PAC placement, there may be small heterogeneity in the time interval between ROSC and start of the study. However, we do not believe that this influenced our results since we studied 24 h hemodynamics and the maximal time delay between ROSC and start of the study is estimated to be around 3 h. Finally, this study was conducted in the hypothermia era. A recent trial failed to show benefit of hypothermia on top of targeted temperature management.15 It’s uncertain whether our suggested hemodynamic targets would still hold in normothermic patients. It has been suggested that the MAP associated with optimal cerebral perfusion may be altered by therapeutic hypothermia.16 In conclusion, we showed that a MAP range between 76–86 mmHg and SVO2 range between 67% and 72% were associated with maximal survival. Based on our hemodynamic model, the optimal cerebral saturation was determined to be 67–68%. This was achieved with a MAP between 87–101 mmHg and a SVO2 between 70% and 75%. Prospective interventional studies are needed to investigate whether forcing MAP and SVO2 in the suggested range with additional pharmacological support would improve outcome. Conflict of interest statement The authors do not have anything to disclose. Acknowledgements The authors would like to thank Annelies Gerits and An Creemers, our consulting statisticians. References 1. Jones AE, Shapiro NI, Kilgannon JH, Trzeciak S, Emergency Medicine Shock Research Network (EMSHOCKNET) investigators. Goal-directed hemodynamic optimization in the post-cardiac arrest syndrome: a systematic review. Resuscitation 2008;77:26–9. 2. Peberdy MA, Callaway CW, Neumar RW, et al. Post-cardiac arrest care: 2010 American Heart Association guidelines for cardiopulmonary resuscitation and emergency cardiovascular care. Circulation 2010;122:S768–86. 3. Sundgreen C, Larsen FS, Herzog TM, Knudsen GM, Boesgaard S, Aldershvile J. Autoregulation of cerebral blood flow in patients resuscitated from cardiac arrest. Stroke 2001;32:128–32. 4. Badeer HS. Contractile tension in the myocardium. Am Heart J 1963;66:432–4. 5. Nolan JP, Soar J. Post resuscitation care – time for a bundle? Resuscitation 2008;76:161–2. 6. Meex I, Dens J, Jans F, et al. Cerebral tissue oxygen saturation during therapeutic hypothermia in post-cardiac arrest patients. Resuscitation 2013;84:788–93. 7. Bouzat P, Suys T, Sala N, Oddo M. Effect of moderate hyperventilation and induced hypertension on cerebral tissue oxygenation after cardiac arrest and therapeutic hypothermia. Resuscitation 2013;84:1540–5. 8. Kilgannon JH, Jones AE, Parrillo JE, et al. Relationship between supranormal oxygen tension and outcome after resuscitation from cardiac arrest. Circulation 2011;123:2717–22. 9. Kilgannon JH, Jones AE, Shapiro NI, et al. Association between arterial hyperoxia following resuscitation from cardiac arrest and in-hospital mortality. JAMA 2010;303:2165–71. 10. Ameloot K, Meex I, Genbrugge C, et al. Accuracy of continuous thermodilution cardiac output monitoring by pulmonary artery catheter during therapeutic hypothermia in post-cardiac arrest patients. Resuscitation 2014;85:1263–8. 11. Beylin ME, Perman SM, Abella BS, et al. Higher mean arterial pressure with or without vasoactive agents is associated with increased survival and better neurological outcomes in comatose survivors of cardiac arrest. Intensive Care Med 2013;39:1981–8. 12. Trzeciak S, Jones AE, Kilgannon JH, et al. Significance of arterial hypotension after resuscitation from cardiac arrest. Crit Care Med 2009;37:2895–903. 13. Kilgannon JH, Roberts BW, Jones AE, et al. Arterial blood pressure and neurological outcome after resuscitation from cardiac arrest. Crit Care Med 2014;42:2083–91. 14. Gaieski DF, Band RA, Abella BS, et al. Early goal-directed hemodynamic optimization combined with therapeutic hypothermia in comatose survivors of out-of-hospital cardiac arrest. Early goal-directed hemodynamic optimization combined with therapeutic hypothermia in comatose survivors of out-ofhospital cardiac arrest. Resuscitation 2009;80:418–24. 15. Nielsen N, Wetterslev J, Cronberg T, et al. Targeted temperature management at 33◦ C versus 36◦ C after cardiac arrest. N Engl J Med 2013;369:2197–206.
62
K. Ameloot et al. / Resuscitation 91 (2015) 56–62
16. Lee JK, Brady KM, Mytar JO, et al. Cerebral blood flow and cerebrovascular autoregulation in a swine model of pediatric cardiac arrest and hypothermia. Crit Care Med 2011;39:2337–45. 17. Paarmann H, Heringlake M, Heinze H, et al. Non-invasive cerebral oxygenation reflects mixed venous oxygen saturation during the varying haemodynamic conditions in patients undergoing transapical transcatheter aortic valve implantation. Interact Cardiovasc Thorac Surg 2012;14:268–72.
18. Taussky P, O’Neal B, Daugherty WP, et al. Validation of frontal near-infrared spectroscopy as noninvasive bedside monitoring for regional cerebral blood flow in brain-injured patients. Neurosurg Focus 2012;32:E2. 19. Bro-Jeppesen J, Annborn M, Hassager C, et al. Hemodynamics and vasopressor support during targeted temperature management at 33◦ C versus 36◦ C after out-of-hospital cardiac arrest: a post hoc study of the target temperature management trial. Crit Care Med 2015;43:318–27.
