Aim:
In a previous study, low and high-normal arterial carbon dioxide tension (PaCO
2
) were not associated with serum neuron specific enolase (NSE) in cardiac arrest survivors. We assessed the effect of PaCO
2
on NSE in cerebrospinal fluid (CSF) and serum.
Methods:
This was a retrospective study. PaCO
2
for the first 24 h was analysed in four means, qualitative exposure state (qES), time-weighted average (TWA), median, and minimum-maximum (Min-Max). These subgroups were divided into low (LCO
2
) and high PaCO
2
(HCO
2
) groups defined as PaCO
2
≤35.3 and PaCO
2
>43.5 mmHg, respectively. NSE was measured at 24, 48, and 72 h (sNSE
24,48,72
and cNSE
24,48,72
) from return of spontaneous circulation (ROSC). The primary outcome was the association between PaCO
2
and the NSE measured at 24 h after ROSC.
Results:
Forty-two subjects (male, 33; 78.6%) were included in total cohort. PaCO
2
in TWA subgroup was associated with cNSE
24,48,72
, while PaCO
2
in the other subgroup were only associated with cNSE
24
. PaCO
2
and cNSE in qES subgroup showed good correlation (r= -0.61;
p<
0.01), and in TWA, Median, and Min-Max subgroup showed moderate correlations (r= -0.57, r= -0.48, and r= -0.60;
p<
0.01). Contrastively, sNSE was not associated and correlated with PaCO
2
in all analysis. Poor neurological outcome in LCO
2
was significantly higher than HCO
2
in qES, TWA, and Median subgroups (
p<
0.01,
p<
0.01, and
p=
0.02).
Conclusion:
Association was found between NSE and PaCO
2
using CSF, despite including normocapnic ranges; TWA of PaCO
2
may be most strongly associated with CSF NSE levels. A prospective, multi-centre study is required to confirm our results.