Cardiac index during therapeutic hypothermia: which target value is optimal?

Mild therapeutic hypothermia is now recognized as standard therapy in patients resuscitated from out-of-hospital cardiac arrest (OHCA), and is recommended in comatose patients suffering from cardiac arrest related to ventricular fibrillation (VF) [1]. In these patients, maintaining an adequate tissue oxygen delivery (DO2) is crucial. However, during hypothermia, clinical signs of hypoperfusion such as cold, clammy skin and delayed capillary refill are not reliable and monitoring devices must, therefore, be used to measure or estimate the cardiac index (CI). However, there are no recommendations regarding the target value of CI in the hypothermic patient. In this article, the authors attempt to provide clinicians with some rationale to guide their therapy for the management of CI in patients treated with mild therapeutic hypothermia.


Introduction
Mild therapeutic hypothermia is now recognized as standard therapy in patients resuscitated from out-ofhospital cardiac arrest (OHCA), and is recommended in comatose patients suff ering from cardiac arrest related to ventricular fi brillation (VF) [1]. In these patients, maintaining an adequate tissue oxygen delivery (DO 2 ) is crucial. However, during hypothermia, clinical signs of hypo perfusion such as cold, clammy skin and delayed capillary refi ll are not reliable and monitoring devices must, therefore, be used to measure or estimate the cardiac index (CI). However, there are no recommendations regarding the target value of CI in the hypo thermic patient. In this article, the authors attempt to provide clinicians with some rationale to guide their therapy for the management of CI in patients treated with mild therapeutic hypothermia.

M ild therapeutic hypothermia
Neurologic outcome and survival rates are improved in patients treated with mild therapeutic hypothermia [2,3]. Th e reason for the improved survival is probably related to the preservation of cerebral function. During mild therapeutic hypothermia, clinical data demonstrate that heart rate is signifi cantly reduced, an eff ect that usually improves left ventricular (LV) fi lling [4] . Whereas CI usually decreases with hypothermia, mild therapeutic hypothermia exerts positive inotropic eff ects in isolated human and pig myocardium. Th e phenomenon of increased inotropism during mild therapeutic hypothermia is not associated with increased sarcoplasmic reticulum Ca 2+content or increased Ca 2+ -transients [5] . Moreover, recent studies using animal species and in humans have provided accumulating evidence suggesting that mild therapeutic hypothermia may also improve cardiac performance [5,6]. Th erefore, the higher survival rates may also be related to positive hemodynamic eff ects of cooling in patients already suff ering from cardiac disease. Furthermore, a study about the hemodynamic eff ects of mild therapeutic hypothermia in 20 consecutive patients admitted in cardio genic shock after successful resuscitation from OHCA showed that these patients seemed to benefi t from mild therapeutic hypothermia in terms of myocardial performance, catecholamine usage, and survival when compared to a historic control group of matched patients treated without hypothermia [7]. Moreover, animal studies have shown that, in myocardial infarction, hypothermia decreases oxygen consumption and in farct size [8]. Th e positive inotropic eff ect of mild therapeutic hypothermia measured by systolic func tion has also been demonstrated in in vivo studies [5,9 ] an d can be measured echocardiographically by the signifi cant increase in ejection fraction (EF) and the augmented contraction velocity measured by pulse contour analysis. However, as shown by Lewis et al., increasing the heartrate (HR) under hypothermic conditions has a neg ative impact on LV contractility [9]. Although systolic performance is clearly improved at all temperature steps investigated, pronounced hypothermia may impair diastolic function [10]. However, in the temperature range recommended for mild therapeutic hypothermia in cardiac arrest patients (32-34°C) [11,12] , dias tolic function seems to be preserved [6]. Al though mild therapeutic hypothermia may have direct repercussions on the myocardium, a study by Bernard et al. showed no clinically signifi cant eff ect on cardiac arrhythmias in the hypothermia group [13].

Wh ich CI target value is optimal in hypothermic patients
In order to maintain perfusion pressure and as a result of hypothermia, the systemic vascular resistance (SVR) increases. As a result, mean arterial pressure (MAP)

R E V I E W
*Correspondence: karim.bendjelid@hcuge.ch Intensive Care Service, Geneva University Hospitals, 4 Rue Gabrielle Perret-Gentil, CH-1211 Geneva 14, Switzerland decreases only slightly with mild therapeutic hypothermia despite a signifi cant decrease in CI. Th is reaction to hypothermia is explained by the vaso constriction of peripheral arteries and arterioles [14] and the stabilization of MAP reducing the vasopres sor dosage. Furthermore, the need for volume in mild ther apeutic hypothermia can be explained by the induction of 'cold diuresis' through a combination of increased venous return (vasoconstriction), activation of atrial na triuretic peptide, decreased levels of antidiuretic hormone and renal antidiuretic hormone receptor levels, and tu bular dysfunction [15,16].
Th ere is considerable confusion about t he s tandard of care for critically ill patients undergoing mild therapeutic hypothermia, in particular related to the subject of hemodynamic optimization. For example, for many years it was commonly taught that although the patient's body temperature decreases, CI values should be as normal as those of normothermic patients. However, although a signifi cant decrease in CI may lead to inadequate organ perfusion in normothermic patients, Bergman et al. failed to demonstrate that a low cardiac output caused lower mixed venous oxygen saturation (SvO 2 ) in patients undergoing mild therapeutic hypothermia [17]. Th is sugges ts that, parallel to the drop in CI, oxygen consumption decreases also because of the lower body temperature. In other words, during mild therapeutic hypothermia, the workload for the heart may be lower because of the lower resting energy metabolism required at a lower body temperature [17]. In fact, the overall metabolic rate decreases by approximately 8 % per °C amounting to a decrease of 32 % when the target tempera ture of 33 °C is reached, thus oxygen consumption and CO 2 production are reduced. Th is eff ect holds true for the heart itself, in which the di minished heart rate reduces the metabolic d emand even further. In addition, mild therapeutic hypothermia induces coronary vasodilata tion and increases myocardial perfusion [18]. Th is belief is corroborated by the fact that under cardiopulmonary bypass (CPB), a technique performed with moderate system ic hypothermia (28 to 32 °C), blood fl ow and CI are maintained between 2.2 and 2.4 l/min/m 2 , without detrimental eff ects.

Bohr eff ects du ring alph a-and Ph-stat hypothermia
During hypothermia, SvO 2 measurement depends on the blood gas analysis technique (alpha-stat, pH-stat) used and the impact of the decrease in temperature on the affi nity of hemoglobin for oxygen (Bohr Eff ect). Indeed, whether a pH-stat or alpha-stat strategy is the ideal acidbase management during severe hypothermic circulatory arrest has been the subject of contention. Advocates of pH-stat management (which aims for a partial pressure of CO 2 [PCO 2 ] of 40 mm Hg and pH of 7.40 at the patient's actual t emperature) claim that the resulting higher CO 2 causes cerebral vasodilatation and faster and more homogeneous cooling. Th ey also suggest that the resulting acidotic protocol of this acid-base management facilitates the release of oxygen from hemoglobin, a fact that off sets the hypothermic leftward shift of the oxygen dissociation curve (Bohr Eff ect). On the other hand, proponents of alpha-stat management, in which there is an alkaline drift during hypothermia, state that this allows cerebral auto-regulation to continue and that cellular transmembrane pH gradients and protein function are maintained. Indeed, when alpha-stat pH manage ment is used, the PCO 2 decreases (and solubility increases); thus a hypothermic patient with a pH of 7.40 and an arterial PCO 2 of 40 mm Hg (measured at 37 °C) will, in reality, have a lower PaCO 2 and this will manifest as a relative respiratory alkalosis coupled with decreased cerebral blood fl ow. In addition, the alkaline pH improves cerebral protection during the ischemic insult. However, there is evidence to suggest that the best technique for acid-base management in patients undergoing deep hypothermic circulatory arrest during cardiac surgery is also dependent upon the age of the patient with better results using alpha-stat in the adult than in the pediatric patient [19].

What exactly does SvO 2 mean during mild therapeutic hypothermia?
Although mild therape utic hypothermia induces a decrease in both HR and CI, in the majority of cases SvO 2 value remains stable (Fig. 1). However, during this condition there is som etimes an increase in systemic arterial lactate levels and it is unclear whether this is caused by increased anaerobic metabolism. Th e pathogenesis of this disorder is uncertain, but it appears not to relate to inadequate DO 2 [20]. Th erefore, the use of inotropic drugs in order to increase CI to 'normal' values may be futile or even harmful because of its negative impact on LV contractility, ventricular arrhythmias and increase in oxygen uptake (VO 2 ). Th e best indicator of good tissue perfusion in patients undergoing mild therapeutic hypothermia seems to be SvO 2 and not CI. However, another misconception arises from the relatively large diff erences between SvO 2 values measured in patients undergoing the alpha-stat and the pH-stat acidbase management. Indeed, oxygen extraction is decreased during mild therapeutic hypothermia as the oxyhemoglobin dissociation curve shifts left (Bohr Eff ect). And, the increased oxygen affi nity of hemoglobin during this hypothermic state could also be aggravated by the alkalotic environment (oxyhemoglobin dissociation curve shifts left; Bohr Eff ect) produced by the alpha-stat method, to the point of developing tissue hypoxia [21]. In this sense, SvO 2 monitoring could be a valuable tool to optimize DO 2 in the hypothermic patient, where the optimal value of SvO 2 is adjusted according to the acidbase management of blood gas measurements.

Conclusion
Although mild therapeutic hypothermia is now recognized as the standard therapy in patients resuscitated from OHCA, optimal target CI values are not clear. However, based on the pathophysiology of the eff ects of hypothermia, it is possible to fi nd answers regarding the hemodynamic management of these patients. Indeed, it seems futile and even dangerous to try to normalize CI to 'normal' values. However, it does seem appropriate to monitor SvO 2 values and arterial lactate levels in these patients, taking into account the impact of hypothermia and acid-base management on the oxyhemoglobin dissociation curve.