Logo: to the web site of Uppsala University

Objective

Malignant middle cerebral artery infarction (MMI) is a severe neurological condition. Decompressive craniectomy (DC) is an established lifesaving surgical treatment. However, the role of neurocritical care with monitoring and management of the intracranial pressure (ICP), pressure reactivity index (PRx), cerebral perfusion pressure (CPP), and optimal perfusion pressure (CPPopt) remain unclear. This study aims to examine the dynamics of these variables post-DC in relation to clinical outcome.

Methods

This retrospective study included 70 MMI patients who underwent DC with ICP monitoring of at least 12 hours and available data of clinical outcome (modified Rankin Scale [mRS] at 6 months). The associations between mRS and cerebral physiology (ICP, PRx, CPP, and ∆CPPopt) was analysed and presented in different outcome heatmaps over the first 7 days following DC.

Results

ICP above 15 mmHg was associated with unfavourable outcome, particularly for longer durations. As PRx exceeded zero, outcome worsened progressively, and values above 0.5 correlated to poor outcome regardless of duration. As CPP dropped below 80 mmHg, there was a transition from favourable to unfavourable outcome. Negative ∆CPPopt, particularly below -20 mmHg, corresponded to unfavourable outcome. In two-variable heatmaps, elevated PRx combined with high ICP, low CPP or negative ∆CPPopt correlated with worse outcome.

Conclusion

Invasive ICP-monitoring may provide prognostic information for long-term recovery in MMI patients post-DC. The study highlighted disease-specific optimal physiological intervals for ICP, PRx, CPP, and ΔCPPopt. Of particular interest, the autoregulatory variable, PRx, influenced the safe and dangerous ICP, CPP, and ∆CPPopt intervals.

OBJECTIVE: There is a paucity of studies on the optimal thresholds for neurointensive care (NIC) targets such as intracranial pressure (ICP) and cerebral perfusion pressure (CPP) in spontaneous intracerebral hemorrhage (sICH). There is also a need to clarify the role of cerebral pressure autoregulatory disturbances (pressure reactivity index [PRx]) and to determine if the autoregulatory CPP target (optimal CPP [CPPopt]) is superior to traditional fixed CPP targets in sICH. In this study, the authors aimed to explore the role of ICP, PRx, CPP, and CPPopt insults in sICH patients treated in the NIC unit.

METHODS: In this observational study, 184 adults with sICH with intracerebral hemorrhage (ICH) volume above 10 ml who received > 12 hours of ICP monitoring during the first 7 days at the authors' NIC unit, Uppsala University Hospital, Sweden, between 2010 and 2019 (10 years) were included. Demographic characteristics, admission status, radiological examination, and clinical outcome were evaluated. Favorable outcome was defined as conscious at discharge, while unfavorable outcome as unconscious or deceased. ICP, CPP, PRx, and CPPopt during the first 7 days were analyzed in relation to outcome.

RESULTS: In total, 138 (75%) patients recovered favorably at discharge. Lower percentage of good monitoring time with ICP above 25 mm Hg was independently associated with favorable outcome. CPP above 80 mm Hg was frequent and independently associated with favorable outcome. Median PRx did not differ between the outcome groups, but there was a trend toward worse outcome when PRx exceeded +0.5. Furthermore, when PRx was analyzed together with the concurrent ICP and CPP values, higher values increased the ICP and CPP interval associated with unfavorable outcome. Lastly, there was no independent correlation between CPP deviation from CPPopt and outcome.

CONCLUSIONS: Avoiding ICP elevations above 20 to 25 mm Hg and maintaining CPP above 80 mm Hg may be beneficial in sICH patients with large bleeding volume who require NIC. PRx was not independently associated with outcome, but higher values appeared to narrow the safe zones of ICP and CPP.

This study compared the roles of extraparenchymal autonomic nervous system (ANS) control of cerebral blood flow (CBF) versus intraparenchymal cerebrovascular autoregulation in 487 patients with aneurysmal subarachnoid hemorrhage (SAH) and 413 patients with traumatic brain injury (TBI). Vasomotion intensity of extraparenchymal and intraparenchymal vessels were quantified as the amplitude of oscillations of arterial blood pressure (ABP) and intracranial pressure (ICP) in the very low frequency range of 0.02–0.07 Hz, or periods of 55–15 sec, computed with a bandpass filter. A version of the pressure reactivity index (PRx-55-15) was computed as the correlation of the filtered waveforms, ABP-55-15 and ICP-55-15. Since ABP-55-15 is measured in the radial artery, any influence of cerebral factors must be mediated by the ANS. ICP-55-15 is measured in the brain and is influenced by intraparenchymal chemical and metabolic factors in addition to the ANS. Patient outcome was assessed using the Extended Glasgow Outcome Score (GOSe). Ten-day mean cerebral perfusion pressure (CPP) was negatively correlated with GOSe in the TBI cohort (R = −0.13, p = 0.01) but positively correlated with GOSe in the SAH cohort, (R = 0.32, p < 0.00001), indicating a much greater dependence on ANS support in the form of elevated CPP in SAH. The optimal CPP range for TBI was 60–70 mmHg, but for SAH it was 110–120 mmHg. The percentage of monitoring time with PRx-55-15 < 0.8, indicating very pressure-active cerebral vessels that resist ANS influence via systemic ABP, is positively correlated with GOSe in the TBI cohort (R = 0.14, p = 0.003), but negatively correlated with GOSe in the SAH cohort (R = −0.10, p = 0.004). The TBI cohort optimal PRx-55-15 for patient outcome was −1.0, while the SAH optimum was 0.3. For the TBI cohort, the correlation of ABP-55-15 amplitude with 10-day mean ICP-55-15 amplitude was 0.29. For the SAH cohort the correlation was 0.51, which is stronger (p = 0.0001). The TBI cohort had a median GOSe of 5 (interquartile range [IQR] 3–7), while SAH had a median of 3 (IQR 3–5), which is worse (p < 0.00001). The higher optimal CPP in patients with SAH, more passive optimal pressure reactivity, and greater dependence of cerebral on systemic vasomotion indicate that they require more active support by the ANS and systemic circulation for CBF than patients with TBI. CBF in patients with TBI is more reliant on cerebrovascular autoregulation based on metabolic demand. This appears to be deficient following SAH, making the heightened ANS support necessary. Although this support is beneficial, it does not fully compensate for the loss of cerebrovascular autoregulation, as reflected in the problems in the SAH cohort with delayed cerebral ischemia and poor outcome.

Background: 

This single-center, retrospective study investigated the outcome effect of the combined intensity and duration of differences between actual cerebral perfusion pressure (CPP) and optimal cerebral perfusion pressure (CPPopt), and also for absolute CPP, in patients with traumatic brain injury (TBI) and aneurysmal subarachnoid hemorrhage (aSAH).

Methods: 

A total of 378 TBI and 432 aSAH patients treated in a neurointensive care unit between 2008 and 2018 with at least 24 hours of CPPopt data during the first 10 days following injury, and with 6-month (TBI) or 12-month (aSAH) extended Glasgow Outcome Scale (GOS-E) scores, were included in the study. ∆CPPopt-insults (∆CPPopt=actual CPP−CPPopt) and CPP-insults were visualized as 2-dimensional plots to highlight the combined effect of insult intensity (mm Hg) and duration (min) on patient outcome.

Results: 

In TBI patients, a zone of ∆CPPopt ± 10 mm Hg was associated with more favorable outcome, with transitions towards unfavorable outcome above and below this zone. CPP in the range of 60 to 80 mm Hg was associated with higher GOS-E, whereas CPP outside this range was associated with lower GOS-E. In aSAH patients, there was no clear transition from higher to lower GOS-E for ∆CPPopt-insults; however, there was a transition from favorable to unfavorable outcome when CPP was <80 mm Hg.

Conclusions: 

TBI patients with CPP close to CPPopt exhibited better clinical outcomes, and absolute CPP within the 60 to 80 mm Hg range was also associated with favorable outcome. In aSAH patients, there was no clear transition for ∆CPPopt-insults in relation to outcome, whereas generally high absolute CPP values were associated overall with favorable recovery.

PURPOSE: Cerebral perfusion pressure (CPP) guidance by cerebral pressure autoregulation (CPA) status according to PRx (correlation mean arterial blood pressure (MAP) and intracranial pressure (ICP)) and optimal CPP (CPPopt = CPP with lowest PRx) is promising but little is known regarding this approach in elderly. The aim was to analyze PRx and CPPopt in elderly TBI patients.

METHODS: A total of 129 old (≥ 65 years) and 342 young (16-64 years) patients were studied using monitoring data for MAP and ICP. CPP, PRx, CPPopt, and ΔCPPopt (difference between actual CPP and CPPopt) were calculated. Logistic regression analyses with PRx and ΔCPPopt as explanatory variables for outcome. The combined effects of PRx/CPP and PRx/ΔCPPopt on outcome were visualized as heatmaps.

RESULTS: The elderly had higher PRx (worse CPA), higher CPPopt, and different temporal patterns. High PRx influenced outcome negatively in the elderly but less so than in younger patients. CPP close to CPPopt correlated to favorable outcome in younger, in contrast to elderly patients. Heatmap interaction analysis of PRx/ΔCPPopt in the elderly showed that the region for favorable outcome was centered around PRx 0 and ranging between both functioning and impaired CPA (PRx range - 0.5-0.5), and the center of ΔCPPopt was - 10 (range - 20-0), while in younger the center of PRx was around - 0.5 and ΔCPPopt closer to zero.

CONCLUSIONS: The elderly exhibit higher PRx and CPPopt. High PRx influences outcome negatively in the elderly but less than in younger patients. The elderly do not show better outcome when CPP is close to CPPopt in contrast to younger patients.

Background: Cerebral perfusion pressure (CPP) is an important target in aneurysmal subarachnoid hemorrhage (aSAH), but it does not take into account autoregulatory disturbances. The pressure reactivity index (PRx) and the CPP with the optimal PRx (CPPopt) are new variables that may capture these pathomechanisms. In this study, we investigated the effect on the outcome of certain combinations of CPP or ΔCPPopt (actual CPP-CPPopt) with the concurrent autoregulatory status (PRx) after aSAH. Methods: This observational study included 432 aSAH patients, treated in the neurointensive care unit, at Uppsala University Hospital, Sweden. Functional outcome (GOS-E) was assessed 1-year postictus. Heatmaps of the percentage of good monitoring time (%GMT) of PRx/CPP and PRx/ΔCPPopt combinations in relation to GOS-E were created to visualize the association between these variables and outcome. Results: In the heatmap of the %GMT of PRx/CPP, the combination of lower CPP with higher PRx values was more strongly associated with lower GOS-E. The tolerance for lower CPP values increased with lower PRx values until a threshold of -0.50. However, for decreasing PRx below -0.50, there was a gradual reduction in the tolerance for lower CPP. In the heatmap of the %GMT of PRx/ΔCPPopt, the combination of negative ΔCPPopt with higher PRx values was strongly associated with lower GOS-E. In particular, negative ΔCPPopt together with PRx above +0.50 correlated with worse outcomes. In addition, there was a transition toward an unfavorable outcome when PRx went below -0.50, particularly if ΔCPPopt was negative. Conclusions: The PRx levels influenced the association between CPP/ΔCPPopt and outcome. Thus, this variable could be used to individualize a safe CPP-/ΔCPPopt-range.

Background

In neurointensive care, increased intracranial pressure (ICP) is a feared secondary brain insult in traumatic brain injury (TBI). A system that predicts ICP insults before they emerge may facilitate early optimization of the physiology, which may in turn lead to stopping the predicted ICP insult from occurring. The aim of this study was to evaluate the performance of different artificial intelligence models in predicting the risk of ICP insults.

Methods

The models were trained to predict risk of ICP insults starting within 30 min, using the Uppsala high frequency TBI dataset. A restricted dataset consisting of only monitoring data were used, and an unrestricted dataset using monitoring data as well as clinical data, demographic data, and radiological evaluations was used. Four different model classes were compared: Gaussian process regression, logistic regression, random forest classifier, and Extreme Gradient Boosted decision trees (XGBoost).

Results

Six hundred and two patients with TBI were included (total monitoring 138,411 h). On the task of predicting upcoming ICP insults, the Gaussian process regression model performed similarly on the Uppsala high frequency TBI dataset (sensitivity 93.2%, specificity 93.9%, area under the receiver operating characteristic curve [AUROC] 98.3%), as in earlier smaller studies. Using a more flexible model (XGBoost) resulted in a comparable performance (sensitivity 93.8%, specificity 94.6%, AUROC 98.7%). Adding more clinical variables and features further improved the performance of the models slightly (XGBoost: sensitivity 94.1%, specificity of 94.6%, AUROC 98.8%).

Conclusions

Artificial intelligence models have potential to become valuable tools for predicting ICP insults in advance during neurointensive care. The fact that common off-the-shelf models, such as XGBoost, performed well in predicting ICP insults opens new possibilities that can lead to faster advances in the field and earlier clinical implementations.

Background

Cerebral perfusion pressure (CPP) management in the developing child with traumatic brain injury (TBI) is challenging. The pressure reactivity index (PRx) may serve as marker of cerebral pressure autoregulation (CPA) and optimal CPP (CPPopt) may be assessed by identifying the CPP level with best (lowest) PRx. To evaluate the potential of CPPopt guided management in children with severe TBI, cerebral microdialysis (CMD) monitoring levels of lactate and the lactate/pyruvate ratio (LPR) (indicators of ischemia) were related to actual CPP levels, autoregulatory state (PRx) and deviations from CPPopt (ΔCPPopt).

Methods

Retrospective study of 21 children ≤ 17 years with severe TBI who had both ICP and CMD monitoring were included. CPP, PRx, CPPopt and ΔCPPopt where calculated, dichotomized and compared with CMD lactate and lactate-pyruvate ratio.

Results

Median age was 16 years (range 8–17) and median Glasgow coma scale motor score 5 (range 2–5). Both lactate (p = 0.010) and LPR (p =  < 0.001) were higher when CPP ≥ 70 mmHg than when CPP < 70. When PRx ≥ 0.1 both lactate and LPR were higher than when PRx < 0.1 (p =  < 0.001). LPR was lower (p = 0.012) when CPPopt ≥ 70 mmHg than when CPPopt < 70, but there were no differences in lactate levels. When ΔCPPopt > 10 both lactate (p = 0.026) and LPR (p = 0.002) were higher than when ΔCPPopt < –10.

Conclusions

Increased levels of CMD lactate and LPR in children with severe TBI appears to be related to disturbed CPA (PRx). Increased lactate and LPR also seems to be associated with actual CPP levels ≥ 70 mmHg. However, higher lactate and LPR values were also seen when actual CPP was above CPPopt. Higher CPP appears harmful when CPP is above the upper limit of pressure autoregulation. The findings indicate that CPPopt guided CPP management may have potential in pediatric TBI.

Purpose: Impaired cerebral pressure autoregulation is common and detrimental after acute brain injuries. Based on the prevalence of delayed cerebral ischemia in aneurysmal subarachnoid hemorrhage (aSAH) patients compared to traumatic brain injury (TBI), we hypothesized that the type of autoregulatory disturbance and the optimal PRx range may differ between these two conditions. The aim of this study was to determine the optimal PRx ranges in relation to functional outcome following aSAH and TBI, respectively.

Methods: In this observational study, 487 aSAH patients and 413 TBI patients, treated in the neurointensive care, Uppsala, Sweden, between 2008 and 2018, were included. The percentage of good monitoring time (%GMT) of PRx was calculated within 8 intervals covering the range from -1.0 to + 1.0, and analyzed in relation to favorable outcome (GOS-E 5 to 8).

Results: In multiple logistic regressions, a higher %GMTs of PRx in the intervals -1.0 to -0.5 and + 0.75 to + 1.0 were independently associated with a lower rate of favorable outcome in the aSAH cohort. In a similar analysis in the TBI cohort, only positive PRx in the interval + 0.75 to + 1.0 was independently associated with a lower rate of favorable outcome.

Conclusion: Extreme PRx values in both directions were unfavorable in aSAH, possibly as high PRx could indicate proximal vasospasm with exhausted distal vasodilatory reserve, while very negative PRx could reflect myogenic hyperreactivity with suppressed cerebral blood flow. Only elevated PRx was unfavorable in TBI, possibly as pressure passive vessels may be a more predominant pathomechanism in this disease.

The aim of this study was to investigate if the absolute pressure reactivity index (PRx) value influenced the association between cerebral perfusion pressure (CPP) and outcome and if the optimal CPP (CPPopt) curve shape influenced the association between deviation from CPPopt and outcome in traumatic brain injury (TBI). We included 383 TBI patients treated at the neurointensive care in Uppsala between 2008 and 2018 with at least 24 h of CPP data. To determine the influence of absolute PRx values on the association between absolute CPP and outcome, the percentage of monitoring time for combinations of CPP and PRx were correlated with outcome (Extended Glasgow Outcome Scale [GOS-E]) in a heatmap. To determine the association between CPP and the relatively best PRx (CPPopt), the percentage of monitoring time of ΔCPPopt (actual CPP-CPPopt) ±5 mm Hg was analyzed in relation to GOS-E. To determine the association between CPP and the relatively best PRx within a certain absolute PRx range (curve shape), both the percentage of ΔCPPopt within the absolute limits of reactivity (PRx <0.00, < 0.15, etc.) and within certain confidence intervals of PRx-deterioration (+0.025, +0.05 etc.) from CPPopt were analyzed in relation to GOS-E. The heatmap of PRx and absolute CPP versus outcome indicated that the CPP range (55-75 mm Hg) associated with favorable outcome was wider when PRx was below 0, whereas the upper CPP-threshold decreased as PRx increased. CPPopt could be calculated during 53% of the monitoring time. Higher percentage of monitoring time with ΔCPPopt ±5 mm Hg, ΔCPPopt within the reactivity-thresholds (PRx <0.30), and ΔCPPopt within the PRx-confidence interval +0.025 were all independently associated with favorable outcome in separate logistic regressions. These regressions had similar area under receiver operating curve and were not superior to a similar regression when the CPPopt-target was replaced by the percentage of monitoring time within the traditional fixed CPP-targets 60 to 70 mm Hg. Individualized CPPopt-targets exhibited a comparable outcome association as traditional CPP targets and different definitions of the best CPPopt range based on the PRx value had a limited effect on the association between deviation from CPPopt and outcome. Since CPPopt could only be calculated during half of the time, an alternative approach would be to assess the absolute PRx to anticipate a safe CPP range