Better lactate clearance associated with good neurologic outcome in survivors who treated with therapeutic hypothermia after out-of-hospital cardiac arrest

We used TH for all OHCA patients after ROSC. Hypothermia induction was started by 4°C cold saline infusion (about 30 ml/kg) over 30 minutes. In addition, we used a commercial temperature regulation system that consisted of hydrogel pads (Arctic Sun Temperature Management System, Medivance, CO, USA). If the commercial device was not available, we used conventional methods such as a commercial cold blanket, 4°C cold saline bladder irrigation or traditional ice packs placed on the patient’s groin, armpits, and around the neck and head. The target temperature for induction was 33.5°C. TH was applied for 24 hours, with the patient’s core body temperature maintained at 33 ± 0.5°C. After the maintenance period, we rewarmed patients to 36.5°C at a rate of 0.15°C/hour on a hydrogel pad or at a rate lower than 0.25°C/hour on a cooling blanket. In the normothermic period after the temperature had reached 36.5°C, the patient’s core body temperature was maintained at less than 37.5°C until 72 hours after ROSC.

During TH, shivering was assessed using the Bedside Shivering Assessment Scale [18]. Shivering was controlled by intermittent intravenous pethidine (25 mg; maximum 100 mg/day) and intravenous magnesium sulfate (2 g intravenously, if serum magnesium <3 mg/dl) at Bedside Shivering Assessment Scale >2 (mild shivering). Neuromuscular blocking agents were used if shivering was not controlled with pethidine or magnesium, according to the decision of the attending physicians, and with electroencephalogram monitoring.

The hemodynamic optimization targets were as follows: central venous pressure ≥8 to 12 mmHg, mean arterial pressure ≥75 mmHg, peripheral oxygen saturation >94%, and partial pressure of carbon dioxide >40 mmHg. To attain these targets, dopamine was used as a first-line vasopressor (starting at 10 μg/kg/minute; maximum 30 μg/kg/minute), and norepinephrine (starting at 5 μg/minute; maximum 100mcg/minute) as a second choice. Thereafter, the use of vasopressin, an intra-arterial balloon pump or extracorporeal membrane oxygenation was determined by the attending physician.

Venous blood samples were collected at the following time points: at TH induction (0 hours), and after 6 hours, 12 hours, 24 hours, 48 hours, and 72 hours. The venous blood samples were stored in sodium fluoride/potassium oxalate tubes. These samples were immediately sent for laboratory testing and centrifuged at 3,500 rpm for 5 minutes to obtain the serum. Lactate levels were then measured by Modular DP analyzer® (Roche, Basel, Switzerland).

The study protocol was reviewed and approved by the hospital Institutional Review Board. In the Samsung Medical Center TH database, data are recorded on standardized case forms following the Utstein style [19, 20] along with the Sequential Organ Failure Assessment score from hypothermia induction. We defined organ failure as Sequential Organ Failure Assessment score >2 for the particular organ.

All adult OHCA patients (>18 years) who underwent TH after ROSC were enrolled into this study. The exclusion criteria for TH were as follows: the cause of arrest was sepsis, progression of malignancy, trauma, or hemorrhagic shock; expected survival was <3 months before cardiac arrest; the patient was capable of only limited self-care and was confined to a bed or chair for 50% or more of waking hours; the patient had a major operation (head, chest, abdomen and vascular) within 7 days; >12 hours had elapsed after ROSC; the patient had an unwitnessed arrest so the time of arrest could not be determined; the patient had an intracranial hemorrhage on brain noncontrast computerized tomography; or the patient was pregnant. Patients who were transferred to our hospital after hypothermia induction at different hospitals were excluded from this study because their blood samples for lactate measurement could not be obtained.

We divided patients into two groups to compare serial lactate levels according to neurologic outcome. Patients in the good neurologic outcome group had Cerebral Performance Category scores of 1 or 2 at 1 month after OHCA. Patients in the poor neurologic outcome group had Cerebral Performance Category scores from 3 to 5 at 1 month [21]. To compare lactate clearance, we identified patients with initial lactate levels >2.5 mmol/l. The 6-hour and 12-hour lactate clearance was calculated as follows:

6-hour lactate clearance (%) = (0-hour lactate – 6-hour lactate) / 0-hour lactate × 100

12-hour lactate clearance (%) = (0-hour lactate – 12-hour lactate) / 0-hour lactate × 100

Table 1 presents the baseline characteristics of the enrolled patients. The demographic data did not show significant differences except for cause of arrest, bystander cardiopulmonary resuscitation, automated external defibrillator shock, no-flow time, and initial rhythms (P <0.01, P <0.01, P <0.01, P <0.01, and P <0.01, respectively). In the poor neurologic group, low-flow time was longer, and hypothermia induction time was shorter although the difference was not statistically significant (P = 0.43 and P = 0.53, respectively) (Table 1).

There was no significant difference in the initial lactate level between the two groups (6.07 mmol/l in the good outcome group vs. 7.13 mmol/l in the poor outcome group, P = 0.42). However, lactate levels at 6 hours, 12 hours, 24 hours, and 48 hours in the good neurologic outcome group were significantly lower than those in the poor neurologic outcome group (3.81 vs. 6.00, P <0.01; 2.91 vs. 5.00, P <0.01; 2.17 vs. 3.86, P <0.01; and 1.57 vs. 2.21, P = 0.03, respectively). Lactate levels at 72 hours in both groups showed no statistically significant difference (1.52 vs. 1.97, P = 0.14) (Figure 2). In addition, for reference, we plotted the median lactate levels of a group of 12 patients who died within 7 days (Figure 2). These levels were higher than both groups of surviving patients in our study.

Patients with initial lactate above 2.5 mmol/l (26 (76.5%) of the good outcome group, 34 (81.0%) of the poor outcome group) were selected as the high lactate subgroup to determine the association between lactate clearance and neurologic outcome. The 6-hour and 12-hour lactate clearance in the good neurologic outcome group were higher than those in the poor neurologic outcome group (35.3% vs. 6.89%, P = 0.01; 54.5% vs. 25.6%, P <0.01), respectively) (Figure 3). In the multivariate regression model, 6-hour and 12-hour lactate clearance showed a significant association with good neurologic outcome after adjustment for confounders (odds ratio: 1.02, 95% confidence interval: 1.03 to 1.05, P = 0.03; and odds ratio: 1.03, 95% confidence interval: 1.00 to 1.07, P = 0.02) (Table 2).

There are several limitations to our study. First, this was a retrospective study using a prospective TH registry data. Since the study was not performed on an intention-to-treat basis, there could have been loss of patients, resulting in selection bias. However, every enrolled patient had been treated in accordance with a prospectively designed protocol and patients’ laboratory findings and major information were retrievable except for three patients. Second, this study was a nonexperimental observational study, which can only determine the association and not causal relationships. We could not fully explain the pathophysiology of our findings. Third, our study was conducted with a relatively small population from a single institute, which also may cause potential bias. However, we can assume that patients had relatively homogeneous baseline characteristics because of the specific geographic location and the similarities between the two groups as shown in Table 1. Fourth, in this study, we did not include the medications and methods used during treatment. The doses of epinephrine used during resuscitation and of vasopressor during TH are known to lead to or aggravate hyperlactemia [34], and thus these could be confounders in the analysis. Fifth, we excluded one patient with underlying liver cirrhosis. Owing to his poor liver function, the patient’s basal lactate level before cardiac arrest was already high and we were concerned this might have a deviating effect on our analysis. We could not perform further analysis due to the small number of patients. Further study might be needed for the group of patients with impaired liver function.