Selected Correlation Dosage als quantitativer Index für Störungen der zerebralen Autoregulation bei Patienten auf der neurochirurgischen Intensivstation.

In neurocritical care, impairment of cerebral autoregulation and reduced intracranial compliance due to brain swelling are frequent and potentially life–threatening events. Selected correlation analysis (SCA) is a method to detect impaired autoregulation combined with severe brain swelling by calculating a correlation-based index sc using fixed-length windows of arterial blood pressure (ABP) and intracranial pressure (ICP) signals. If sc is higher than a predefined significance level and the correlation is positive, the windows are labeled scp, indicating the pathophysiology mentioned above. We examined whether the quantitative extent of sc under scp conditions, represented by the area under the curve of sc for scp segments, can serve as a useful continuous parameter, indicating the severity of autoregulation failure.

We included 77 patients (46 female and 31 male) with a median age of 50 years, treated for subarachnoid hemorrhage (SAH; n = 58) or traumatic brain injury (TBI; n = 19). The median initial GCS was 8, and the median GOS at the last follow-up was 3. SCA was used to extract all segments of scp (sc > 0.0556), and for these segments, the sc dosage was calculated by area under the curve (AUC) of the sc values. The resulting value for one segment is called scp_dosage of this segment. In the first step we calculated the sum of scp_dosage per patient scaled by the measurement length and analyzed whether there exists, in analogy to the scp values, a significant correlation to the GOS. Additionally, we selected the maximum, minimum, and mean values of scp_dosage per patient and repeated the above-mentioned statistical analysis with these values.

We could demonstrate that the scp_dosage per patient scaled by the measurement length strongly correlates to patient outcome after the last follow-up with a correlation coefficient of -0.4155 and a p-value of 0.0002. This also holds for the maximum of the scp_dosage per patient, leading to a correlation coefficient of -0.348 and a p-value of 0.0019. The correlation analysis of the mean value of scp_dosage per patient leads to a correlation coefficient of -0.384 and a p-value of 0.0006.

scp_dosage and some of its variations are valid outcome predictors. The maximum value of scp_dosage is of great interest as this value might lead to a limit value of scp_dosage that should not be reached during treatment.