• Citer

Subarachnoid Hemorrhage (SAH): from epidemiology to neuromonitoring

par Gouvêa Bogossian, Elisa
Président du jury Verhoeven, Caroline
Promoteur Taccone, Fabio
Co-Promoteur Creteur, Jacques
Publication Non publié, 2022-03-07

Thèse de doctorat

• Accès en ligne
• Détails
• Contenu
• Statistiques

Résumé :

Introduction: Subarachnoid hemorrhage (SAH) is a rare but devastating form of stroke, responsible for elevated mortality and disability rates. The pathophysiology of brain injury in SAH can be divided into two distinct phases: early brain injury (EBI) and delayed cerebral ischemia (DCI), both being important determinants of unfavorable outcome (UO). Secondary brain injury caused by systemic events is also a relevant mechanism that aggravates prognosis. To better assess SAH patients at admission, different clinical scales were developed. The world federation of neurological surgeons (WFNS) scale classifies patients according to their Glasgow coma scale (GCS) and motor deficit at presentation. Patients with a WFNS of 4 (GCS 7-12) or 5 (GCS ≤6) have worse outcome and are called poor grade patients. This classification can also be used to identify EBI. While clinical examination is the cornerstone of the assessment of SAH patients, it may not always be reliable when they are unconscious or sedated. In this setting multimodal neuro-monitoring (including invasive intracranial pressure – ICP monitoring, brain tissue oxygenation- PbtO2 monitoring, electroencephalogram- EEG, transcranial Doppler and neuroimaging) can help identify neurological deterioration, the need for interventions and the response to them. Objectives: The aims of this thesis are: a) to evaluate the temporal changes in mortality and neurological outcome of poor grade SAH patients, to study the impact of hydrocephalus and intraparenchymal hemorrhage in the outcome of such patients and to study the differences between WFNS score of 4 and 5; b) to evaluate the impact of systemic derangements (i.e. infection, lymphopenia, anemia and hyperlactatemia) as potential causes or biomarkers of secondary brain injury; c) to study the impact of red blood transfusion (RBCT) on brain tissue oxygenation; d) to assess the impact on outcome of the implementation of a combined PbtO2/ICP management strategy compared to ICP only guided therapy.Methods: We conducted 6 single-center retrospective cohort studies of non-traumatic SAH and one multicentric retrospective study of acute brain injury patients. All studies were carried out in accordance with the “strengthening the reporting of observational studies in epidemiology” (STROBE) statement. Descriptive statistics and multivariable analysis adjusting for cofounders were carried out in all studies. Whenever relevant to the study objective, survival analysis was also performed.Results: In the first study (n=353), mortality tended to decrease in 2008-2011 and 2016-2018 periods (HR 0.55 [0.34-0.89] and HR 0.33 [0.20-0.53], respectively, when compared to 2004-2007) in poor grade SAH patients, while the proportion of patients with UO remained high and did not vary significantly over time. Patients with WFNS 5 had higher mortality (68% vs 34%, p=0.001) and more frequent UO (83% vs 54%, p=0.001) than those with WFNS 4. In the multivariable analysis, WFNS 5 was independently associated with mortality (HR 2.12 [1.43-3.14]) and UO (OR 3.23 [1.67-6.25]). The presence of hydrocephalus was associated with a lower risk of mortality (HR 0.60 [0.43-0.84]). In the second study (n=248), 70 (28.2%) developed at least one infection; the most frequent sites of infection were respiratory (57.1%), primary bloodstream (16%) and urinary tract infections (15.7%). Twenty-eight patients (11.3 % of all patients) had at least one episode of septic shock. Infected patients had a higher UO rate (60.0% vs. 33.3%, p=0.001). Diabetes mellitus [SHR 1.79 (CI 95% 1.03-3.13)] and intracranial hypertension [SHR 1.92 (CI 95% 1.14-3.25)] were independently associated with the occurrence of infections. Septic shock [OR 6.36 (CI 95% 1.24-32.51), p=0.02] was independently associated with UO. In the third study (n=270), 121 (45%) patients had lymphopenia and 62 (23%) patients developed infections. The median lymphocyte count on admission was 1280 (890-1977)/mm3. Lymphopenia patients had more episodes of infections (38/121, 31% vs. 24/139, 17% - p= 0.003), while mortality and UO were similar when compared to non-lymphopenia patients. Lymphopenia was not independently associated with the development of infection, UO nor with mortality. In the fourth study (n=456), 158 SAH patients (35%) died in hospital and 209 (46%) had UO at 3 months. The median highest lactate concentration on admission was 2.7 [1.8-3.9] mmol/L. Non-survivors and patients with UO had significantly higher lactate concentrations compared to other patients. Hyperlactatemia (lactate > 2mmol/L) increased the chance of dying (OR 4.19 [95% CI 2.38-7.39]) and of having UO in 3 months (OR 4.16 [95%CI 2.52-6.88]), after adjusting for confounding factors. In the fifth study (n=270), patients with UO had lower hemoglobin (Hb) over time and received RBCT more frequently than others (15/109, 14% vs. 6/161, 4% - p<0.01). Pre-RBCT median Hb values were similar in UO and FO patients (6.9 [6.6-7.1] vs. 7.3 [6.3-8.1] g/dL – p=0.21). The optimal discriminative Hb threshold to predict UO was 9 g/dL. In a multivariable analysis, neither anemia nor RBCT were independently associated with UO. In the sixth study (n=69), we evaluated a total of 109 RBCTs after a median of 9 [5-13 days] days after injury. Baseline Hb and PbtO2 were 7.9 [7.3-8.7] g/dL and 21 [16-26] mmHg, respectively; 2 hours after RBCT, the median absolute Hb and PbtO2 increases from baseline were 1.2 [0.8-1.8] g/dL (p =0.001) and 3 [0-6] mmHg (p=0.001). A 20% increase in PbtO2 after RBCT was observed in 45 (41%) transfusions. High heart rate (HR) and low PbtO2 at baseline were independently associated with a 20% increase in PbtO2 after RBCT. Baseline PbtO2 had an area under the receiver operator characteristic curve of 0.73 (95% CI 0.64-0.83) to predict post- transfusion PbtO2 increase; a PbtO2 of 20 mmHg had a sensitivity of 58% and a specificity of 73% to predict PbtO2 increase after RBCT. The response to RBCT in patients with SAH was similar to other patients (i.e. with traumatic brain injury or intracerebral hemorrhage). In the seventh study (n=163), 62 patients were monitored with PbtO2 and 54 (87%) had at least one episode of brain hypoxia (PbtO2<20mmHg). In patients that required treatment based on neuromonitoring strategies, PbtO2-guided therapy (OR 0.33 [CI 95% 0.12-0.89]) compared to ICP-guided therapy had a protective effect on neurological outcome at 6 months.Conclusions: Hospital mortality and UO rates remain high in poor grade SAH patients; patients with WFNS 5 on admission have worse prognosis than others. Infections in SAH patients are prevalent, especially pneumonia. Septic shock is associated with poor neurological outcome in this group of patients. Early lymphopenia is also common after SAH; however, it is not significantly associated with the development of infections nor with poor outcome. Initial blood lactate concentrations have prognostic implications in patients with SAH. Anemia is frequent after SAH; however, its prognostic value remains to be elucidated. Lower PbtO2 values and tachycardia at baseline could predict a significant increase in brain oxygenation after RBCT. Moreover, combined PbtO2/ICP-guided therapy might be associated with improved long-term neurological outcome, when compared to ICP-guided therapy.