Background: Temperature is an important modulator of the evolution of ischemic brain injury--with hypothermia lessening and hyperthermia exacerbating damage. We recently reported that children resuscitated from predominantly asphyxial arrest often develop an initial spontaneous hypothermia followed by delayed hyperthermia. The initial hypothermia observed in these children was frequently treated with warming lights which, despite careful monitoring, often resulted in overshoot hyperthermia. We have previously reported in a rat model of asphyxial cardiac arrest that active warming, to prevent spontaneous hypothermia, worsens brain injury.
Objective: We sought to determine whether delayed induction of hyperthermia would worsen brain injury after asphyxial arrest in rats.
Design: Male Sprague-Dawley rats were asphyxiated for 8 mins and resuscitated. An implantable temperature probe was placed into the peritoneum before asphyxia. The probe is a component of a computer-based, radiofrequency, telemetry system (Minimitter, Sunriver, OR) that allowed continuous acquisition and manipulation (via heating and cooling devices) of core (intraperitoneal) body temperature. Body temperature was monitored but not manipulated for the first 24 hrs of recovery. Rats were assigned to: no temperature manipulation (n = 21), induced hyperthermia (40 +/- 0.5 degrees C) for 3 hrs beginning at 24 hrs (n = 21), or induced hyperthermia at 48 hrs (n = 10). Control groups included sham rats (all surgical procedures except asphyxia) treated with induced hyperthermia at 24 hrs (n = 4) or 48 hrs (n = 4) and naïve rats (n = 4). Rats were killed at 7 days and injured neurons in hematoxylin and eosin stained coronal brain sections through dorsal hippocampus were scored in a semiquantitative manner on a scale of 0 to 10 (0 = normal; 1 = up to 10% neurons with ischemic neuronal changes; 10 = 90-100% neurons with ischemic neuronal changes). Normal-appearing neurons were also counted in CA1. The number of normal-appearing neurons in a 20x field in CA1 were also counted.
Main results: All naïve and sham hyperthermia control rats survived the protocol. There was a trend toward a larger mortality rate in asphyxiated rats treated with induced hyperthermia at 24 hrs (9 of 21 died) vs. asphyxiated rats without induced hyperthermia (3 of 21) or with hyperthermia induced at 48 hrs (3 of 10) (Kaplan-Meier p=.0595). Asphyxiated rats with hyperthermia induced at 24 hrs had larger (worse) histopathology damage scores than rats subjected to asphyxia without induced hyperthermia (9.3 +/- 1.5 vs. 6.2 +/- 2.6; p=.001). Histopathology damage scores in asphyxiated rats with hyperthermia induced at 48 hrs did not differ from those in rats asphyxiated without induced hyperthermia (6.4 +/- 3.0 vs. 6.2 +/- 2.6; p=.907). There were fewer normal-appearing CA1 neurons in asphyxiated rats with hyperthermia induced at 24 hrs vs. rats subjected to asphyxia without induced hyperthermia (33 +/- 13 vs. 67 +/- 36; p=.002). The number of normal-appearing CA1 neurons in asphyxiated rats with hyperthermia induced at 48 hrs did not differ from that in rats asphyxiated without induced hyperthermia (59 +/- 21 vs. 67 +/- 36; p=.885).
Conclusions: Induced hyperthermia when administered at 24 hrs, but not 48 hrs, worsens ischemic brain injury in rats resuscitated from asphyxial cardiac arrest. This may have implications for postresuscitative management of children and adults resuscitated from cardiac arrest. The common clinical practice of actively warming patients with spontaneous hypothermia might result in iatrogenic injury if warming results in hyperthermic overshoot. Avoidance of hyperthermia induced by active warming at critical time periods after cardiac arrest may be important.