Pfeifer R, Franz M, Figulla HR. Hypothermia after cardiac arrest does not affect serum
levels of neuron-specific enolase and protein S-100b. Acta Anaesthesiol Scand 2014;58:1093-100.
Therapeutic hypothermia (TH) is a treatment option in comatose survivors of out-of-hospital
cardiac arrest (CA) after ventricular fibrillation.[1] The patients are mechanically ventilated and receive sedative drugs for several
days making clinical neurological examination difficult. The predictors of neurological
outcome such as brain-derived proteins neuron-specific enolase (NSE) and protein S-100b
are needed. They are widely used as prognostic markers for evaluating the severity
of hypoxic brain injury in comatose survivors of CA.[2]
The present study included 201 comatose adult patients who had suffered in or out
of hospital nontraumatic CA. They were admitted to intensive care unit (ICU) between
January 2003 and June 2010. The patients who reacted adequately during 1 h observation
phase without sedation and those who died during the first 48 h after CA were excluded
from the study. TH was applied to 140 patients (30 female, 110 male, mean age −63.2
± 14.4 years), whereas 61 (24 female, 37 male, mean age −67.8 ± 14.2 years) received
comparable intensive care therapy without hypothermia. TH was applied according to
criteria published by authors in resuscitation,[3] but later adapted according to the recommendation of European Resuscitation Council.[1]
[4] Serum levels of NSE and S-100b were assessed first immediately on ICU admission
and then subsequently until 5 days after restoration of spontaneous circulation (ROSC)
using commercial immunoluminometric assay with a LIAISON analyzer. The reference values
were 12.5 ng/ml for NSE and 0.15 mcg/L for S-100b. The levels did not influence therapeutic
decisions during the first 7 days after ROSC.
Therapeutic hypothermia was applied either by surface cooling or intravascular cooling
to maintain body core temperature (measured with a bladder temperature probe) within
a narrow range 33 ± 0.5°C for 24 h. Rewarming was achieved passively in surface cooled
patients and actively (0.3°C/h) in intravascular cooled patients. After 4 weeks, the
patients were classified according to Pittsburgh Cerebral Performance Category Scale
(CPC).[5] The patents were divided into two groups; those with poor outcome, that is, those
who died or remained comatose (CPC 4 and 5) and those who had good to moderate neurological
recovery that is, recovery of at least cognitive brain function and survival with
neurological disabilities of variable severity or without any neurological impairment
(CPC 3-1).
In the hypothermia group (HG), 61 of 140 patients (43.6%) and in the normothermia
group (NG) 26 of 61 patients (42.6%) survived the first 4 weeks after ROSC with moderate
to good neurological outcome. Due to adhering to the guidelines for TH, HG patients
were significantly younger than NG patients (P = 0.012), and CA occurred more frequently in out-of-hospital area. The mean application
time of hypothermia amounted to 23.9 h and the target body core temperature (32.5–33.5°C)
was maintained for 16.9 h. The duration of re-warming to a core temperature of 36°C
averaged 9.1 h. A significantly higher level of NSE and S-100b were observed on day
4 and 5 respectively in HG when compared to NG. By trend, the serum levels of both
proteins were lower in the NG.
An intra-aortic balloon pump had to be applied to 38 (27%) of HG and 11 (18%) of NG
patients. For NG and HG patients with unfavorable neurological outcome (CPC 4 and
5), on each day significantly higher NSE and S-100b were found when compared with
those with CPC 3-1. Taking groups together, 87 patients survived with CPC 3-1; in
several patients NSE levels exceeded the threshold of 33 ng/ml. On the 3rd day, NSE levels ≥40 ng/ml predicted poor neurological outcome with a sensitivity
74.1% (confidence interval [CI] = 0.6475–0.8203), a specificity 95.2% (CI = 0.8825–0.9869),
a positive predictive value (PPV) of 95.2% (CI = 0.8825–0.9869). Whereas, S-100b levels
≥1.03 mcg/L on the 3rd day had sensitivity 57.8% (CI = 0.4544–0.6939), specificity 95.6% (CI = 0.8782–0.9909),
PPV 93.2% (CI = 0.8134–9857) for poor neurological outcome.
The improved survival rate in HG patients was not observed, which was also the case
in the previous trial published in 2013.[6] There was a tendency for the serum levels of both proteins to be higher in HG patients.
Lower NSE levels have been reported earlier in HG patients but in those studies more
patients had favorable neurological outcome.[7] The authors consider it unlikely that the kinetics of the two proteins were changed
by TH. There is increasing evidence that resuscitated patients with NSE concentration
much higher than the cut-off level can survive with moderate or good neurological
outcome. Hemolysis or several forms of cancer may influence levels of NSE; similarly,
S-100b is released from tissues other than brains such as adipocytes, chondrocytes
and several forms of cancer of central nervous system and melanoma.
The authors conclude that TH has no influence on NSE and S-100b serum levels in comatose
CA survivors. The increase in both the proteins indicate poor neurological outcome;
hence, their measurement is an additional tool for making prognosis on comatose CA
survivors. However, at present it is not possible to recommend reliable threshold
protein concentration, further investigations in this field are warranted.
