Effect of HHH-Therapy on Regional CBF after Severe Subarachnoid Hemorrhage Studied by Bedside Xenon-Enhanced CT

Patients diagnosed with severe spontaneous SAH admitted to Uppsala University Hospital during the time period 2013–2016 were prospectively enrolled in the study. The diagnosis of SAH was verified by CT. As ventilator treatment is a prerequisite for bedside XeCT in our setting, only patients requiring intubation due to their neurological state at admission or neurological deterioration day 0–1 were included, i.e., patients with more severe SAH. Exclusion criteria were severe intracranial hypertension, deep sedation with thiopental, respiratory problems requiring FiO₂ >0.6, the absence of informed consent, and futility or “do not resuscitate” order.

According to our standard NIC protocol, CBF measurements were scheduled for three time intervals: day 0–3, 4–7, and 8–12. Patients with clinical diagnose of DCI received HHH-therapy as described below. For inclusion in this study, a XeCT measurement within 0–48 h before the start of HHH-therapy was required (used as baseline), and the intention was to perform the next XeCT measurement during ongoing therapy when logistically possible. Data were also collected for non-DCI patients with corresponding XeCT measurements in the same time-windows as a reference group for comparison.

The XeCT procedures in our NIC unit are conducted as described by Gur et al. and Yonas et al. [11‐14], using inhaled xenon as an inert contrast agent during the CT scans. Since xenon is radiopaque, lipid soluble, and crosses the blood brain barrier, its concentration in the cerebral tissue can be measured by CT. Blood flow calculation is based on Kety’s [15, 16] application of the Fick principle; blood flow is proportional to inert gas uptake in tissue. Mobile equipment allows bedside measurements in the NIC unit, which facilitates the clinical use of XeCT [17]. Stable (non-radioactive) xenon at a concentration of 28% is administered to the patients breathing circuit for 4.5 min using the Enhancer 3000 and computer software (Diversified Diagnostic Products Inc, Huston, USA). CT scans during the xenon inhalation are obtained by the CereTom (Neurologica, Boston, USA). The xenon delivery and CT scans are synchronized by the computer software, and the resulting radiologic tissue enhancement of the xenon wash-in enables CBF to be calculated (ml/100 g/min) and plotted as color-coded maps in scans at four different levels of the brain (eight scans per level, two at baseline and six enhanced during xenon wash-in). Levels with extensive artifacts from bone structures or coils/clips were excluded from calculation. Typically, three scan levels could be used for further calculations in each patient. Mean blood flow in each of twenty evenly distributed cortical regions of interest (ROIs) was calculated (Fig. 1) for every level [13], resulting in a total of sixty ROIs. Typical area of each ROI was 350–450 mm². ROIs in areas of hematoma or with artifacts from ventricular drains were excluded.

Management of all patients followed our standard NIC procedures for SAH as earlier described by Ryttlefors et al. [5], where the cornerstones are continuous multimodal monitoring of physiological and biochemical parameters and cautious clinical observation for avoidable factors with prompt interventions to minimize secondary brain injury. Sedation with propofol (Propolipid^®, Fresenius Kabi AB, Uppsala, Sweden) is titrated in the interval 0–4 mg/kg/h, and morphine (Morfin, Meda AB, Solna, Sweden) is administered as needed. Patients with suspicion of hydrocephalus or altered level of consciousness receive a ventriculostomy catheter for monitoring of intracranial pressure (ICP) and drainage of cerebrospinal fluid. Elevated ICP (>20 mmHg) is treated with open ventricular drainage against a pressure level of 15 mmHg. The general policy is to treat aneurysms early with endovascular coil embolization as first choice when possible or surgical clipping. After the aneurysm repair, patients are kept mildly hypervolemic if not contraindicated by intracranial mass effect or elevated ICP. The volume status is maintained by fluid administration in the higher normal range with addition of albumin infusion (Flexbumin^® 200 mg/ml, Baxter AG, Vienna, Austria) if needed and monitored by central venous pressure, clinical evaluation and rigorous fluid balance calculation. Nimodipine (Nimotop^®, Bayer Pharma AG, Berlin, Germany) is given from day one and onwards.

To detect DCI, patients were monitored for general and focal neurological changes by repeated wake-up tests [18]. In the case of neurological deterioration, CT and laboratory tests were performed to rule out intracranial mass, manifest infarction, hydrocephalus, and meningitis. If DCI was concluded in the absence of other causes of deterioration, HHH-therapy to augment CBF was initiated for 5 days if not contraindicated by the intracranial situation, congestive heart failure, fluid overload or severe respiratory problems. Patients received daily infusions of 500 ml dextran solution (Rheomacrodex^®, Meda AB, Solna, Sweden) and 200 ml albumin (Flexbumin^® 200 mg/ml, Baxter AG, Vienna, Austria) and were kept in supine position. In our standard NIC protocol, emphasis is on avoiding severe side effects and the blood pressure targets during HHH-therapy are moderate. The systolic blood pressure (SBP) was closely monitored and kept above 140 mmHg using vasoactive agents if needed. As first line, dobutamine (Hospira, Lake Forest, IL, USA) was used for inotropy or, if insufficient, norepinephrine (Noradrenalin, Hospira Nordic AB, Stockholm, Sweden) was added as vasopressor. The clinical effect of the HHH-therapy was repeatedly evaluated.

Global cortical CBF (ml/100 g/min) was calculated as mean of all ROIs at all scan levels used (Fig. 1). Regional CBF for each vascular territory was calculated as mean of the corresponding ROIs at all scan levels used (Fig. 1): right anterior cerebral artery—ROI 1–2, right middle cerebral artery—ROI 3–8, right posterior cerebral artery—ROI 9–10 and equally for ROI 11–20 on the left side. The vascular territory with lowest CBF was identified as the worst vascular territory in each patient.

To detect and quantify the extent of areas with low blood flow and near-ischemic flow, thresholds for local CBF were set to 20 and 10 ml/100 g/min [19, 20]. The percentage of cortical ROI area with local CBF below these thresholds was then calculated as sum of ROI area (mm²) with local CBF below the threshold divided by the total analyzed ROI area in each patient.

Cerebral infarcts visible on follow-up CT at 12 days or later after admission to the NIC unit were categorized as <20 mm, 20–39 mm, >40 mm, or multiple.

SPSS statistics 23.0 software (IBM Corp, Armonk, NY, USA) was used for statistical analyses of the collected data. Differences in systemic physiological parameters between measurements were tested by paired samples t test. CBF data for groups of patients are presented as median values and interquartile range because of non-normal distribution. Differences in CBF parameters between related samples were tested by Wilcoxon signed-ranks test, and between independent samples by Mann–Whitney U test. Statistical significance level was set at P < 0.05.

Hemodynamic parameters and ventilation were clinically stable during the XeCT procedures. The alterations from start to end in all procedures (N = 96) were small; mean arterial pressure (MAP) mean 92.5 mmHg (CI 90.0–95.0) versus 89.9 (CI 87.6–92.2), ICP mean 14.3 mmHg (CI 13.2–15.4) versus 14.4 (CI 13.5–15.4), PCO₂ mean 5.41 kPa (CI 5.29–5.53) versus 5.51 (CI 5.40–5.63). No adverse events were observed in relation to the procedures.

SBP, MAP, cerebral perfusion pressure (CPP), pCO₂, and hematocrit at start of the XeCT measurements at the different time-windows for the DCI and non-DCI groups are presented in Table 2. The mean SBP was slightly elevated during HHH-therapy in the DCI group, from 151.2 mmHg (CI 142.1–160.3) to 157.3 mmHg (CI 150.7–163.8), but the difference did not reach statistical significance. Among the DCI patients, hematocrit was significantly reduced during HHH-therapy compared to baseline (Table 2).

The use of inotropes and vasopressors was low in all groups. The number of patients and dose range for each group are also presented in Table 2.

Results of the XeCT measurements are presented in Table 2 and Fig. 3. Global cortical CBF at baseline was significantly lower in the DCI group compared to the non-DCI group; median 29.5 ml/100 g/min (IQR 24.6–33.9) versus 34.9 (IQR 29.0–41.7) (P = 0.005).

During HHH-therapy, the DCI group showed an increase in median global cortical CBF from 29.5 (IQR 24.6–33.9) to 38.4 (IQR 27.0–41.2) ml/100 g/min (P = 0.001), while no significant change over time was seen in the non-DCI group. No statistical analyses were performed for differences between DCI patients during HHH-therapy and non-DCI patients at the corresponding time-window.

At baseline, there were also significant differences in regional CBF parameters between the DCI and non-DCI groups (Table 2; Fig. 3); median rCBF of the worst vascular territory for the DCI group was 19.6 (IQR 15.0–24.2) compared to 27.2 (20.8–35.4) ml/100 g/min for the non-DCI group (P = 0.005). Concerning regions with low or near-ischemic CBF, the DCI group had a larger proportion of ROI area with blood flow below 20 ml/100 g/min (median 26.2 vs 11.9%, P = 0.005) and flow below 10 ml/100 g/min (median 6.7 vs 0.7%, P = 0.023) compared to the non-DCI group.

During HHH-therapy, median rCBF of the worst vascular territory for the DCI group increased from 19.6 (IQR 15.0–24.2) to 27.3 (IQR 17.8–34.1) ml/100 g/min (P = 0.006) (Table 2; Fig. 3), and the proportion of ROI area with local blood flow below the threshold of 20 ml/100 g/min was reduced from median 26.2% at baseline to 8.55% (P = 0.019). The area with flow below 10 ml/100 g/min was small already at baseline but tended to decrease (from 6.7 to 0.0%, P = 0.056). The relative blood flow in the worst vascular territory compared to the best hemispheric blood flow (CBF index of worst vascular territory) showed a small but statistically significant increase during HHH-therapy. In the non-DCI group, there were no significant changes from baseline to the second time-window regarding rCBF in the worst vascular territory or the proportion of ROI area with local blood flow below the specified thresholds. In this group, there was a small, but statistically significant reduction in the CBF index of worst vascular territory between the two measurements.

Short-term outcome parameters are presented in Fig. 4. The proportion of patients with favorable clinical course outcome was 65% in the HHH-treated DCI group and 57% in the non-DCI group. The proportion of patients where no infarcts >20 mm were detected at follow-up CT was 65% in the DCI group and 46% in the non-DCI group. Since the non-DCI group is not considered a valid control group for HHH-therapy, no statistical analyses were performed on the differences in outcome parameters.