A high number of disabilities, resource consumption, and deaths result from ischemic stroke (Powers et al., 2018). Oxidative stress can produce secondary brain injury in ischemic stroke, and antioxidants are generated to avoid cell damage due to reactive oxygen species (Manzanero et al., 2013; Olmez and Ozyurt, 2012; Pradeep et al., 2012; Radak et al., 2014; Rodrigo et al., 2013; Warner, 2004). The different antioxidant molecules act synergistically, and determining total antioxidant capacity (TAC) could provide better information on antioxidant status than the determination of each antioxidant compound separately (Ghiselli et al., 2000).

There is scarce data on blood TAC concentrations and prognosis of patients with ischemic stroke (Fernández-Gajardo et al., 2019; Ghonimi et al., 2019; Guldiken et al., 2009; Lorente et al., 2016; Lorenzano et al., 2018, 2019). One study found lower blood TAC concentration measured on the first day of ischemic stroke in patients with poor functional prognosis (Ghonimi et al., 2019). In contrast, higher blood TAC concentrations have been found on the first day of ischemic stroke patients than in healthy subjects (Guldiken et al., 2009), and in non-surviving patients than in surviving ones (Lorente et al., 2016). Another study found no association between blood TAC concentrations during the first week of ischemic stroke and the infarct volume (Fernández-Gajardo et al., 2019). On the other hand, no significant differences were found in plasma TAC concentrations within 9 hours of stroke onset between patients with infarct growth and those without (Lorenzano et al., 2018), while statistically significant differences were observed in plasma TAC concentrations within 9 hours of stroke onset between patients with ischemic penumbra (perfusion-weighted/diffusion-weighted mismatch $>$ 20%) and those without (Lorenzano et al., 2019). Therefore, our study aims to determine whether blood TAC levels during the first seven days of a cerebral infarction can help in mortality prediction.

We analyzed 34 non-surviving and 34 surviving patients. We found higher GCS and platelets in surviving patients than in the non-survivors (Table 1). We found no statistically significant differences in 30-day survival rate in patients without decompressive craniectomy (48%, 25 of 52 patients) or with it (56%, 9 of 16 patients) ($P$ = 0.78), nor in patients without intravenous thrombolysis (50%, 23 of 47 patients) or with it (52%, 11 of 21 patients) ($P$ = 0.99). No patients underwent endovascular thrombectomy. A total of 14 patients died after cardiac arrest, 2 after nosocomial pneumonia, and 18 patients to brain death.

Higher serum TAC concentrations on days one ( 0.001), four ( 0.001), and eight ($P$ = 0.003) were found in non-surviving patients in comparison to surviving ones (Fig. 1). No significant differences were found in serum TAC concentrations at days one, four, and eight within the groups of surviving ($P$ = 0.37) and non-surviving ($P$ = 0.26) patients.

Fig. 1.

Serum total antioxidant capacity (TAC) at days 1${}^{s\mathbf{}t}$, 4${}^{t\mathbf{}h}$ and 8${}^{t\mathbf{}h}$ in survivor and non-survivor patients.

To predict the mortality at 30 days serum TAC concentrations on days one, four and eight of the MMCAI had an area under the curve (95% CI = confidence intervals) of 0.86 (0.75-0.93; 0.001), 0.87 (0.74-0.95; 0.001) and 0.79 (0.64-0.90; $P$ = 0.004)), respectively. Table 2 showed positive and negative likelihood ratios, sensitivity, specificity, positive, and negative predicted values of serum TAC concentrations cut-offs on days one, four, and eight of MMCAI for predicting mortality.

Regression analyses showed an association of serum TAC levels with 30-day mortality controlling for platelet count, GCS and lactic acid (Odds ratio = 2.672; 95% CI = 1.489-4.796; $P$ = 0.001), and also controlling for sex, glycemia and age (Odds ratio = 1.969; 95% CI = 1.288-3.011; $P$ = 0.002) (Table 3).

We found an association between serum concentrations of TAC and malondialdehyde at days one ($\rho$ = 0.46; $P\le$ 0.001), four ($\rho$ = 0.49; $P\le$ 0.001), and eight ($\rho$ = 0.65; $P\le$ 0.001) of a severe MMCAI.

The potential use of serum concentrations of TAC during the first week of severe MMCAI for predicting mortality was the main novel result in our study. On the one hand, no association was found between infarct growth and plasma TAC concentrations within 9 hours of stroke onset (Lorenzano et al., 2018), and between infarct size and blood TAC concentrations during the first week of ischemic stroke (Fernández-Gajardo et al., 2019); however, the association between blood TAC concentrations and mortality was not studied. Also, another study found lower blood TAC concentration on day 1 of ischemic stroke in patients with poor functional prognostic (Ghonimi et al., 2019). On the other hand, an association between high plasma TAC concentrations within 9 hours of stroke onset and brain ischemic penumbra has been found (Lorenzano et al., 2019). Previously, we had determined blood TAC concentration on the first day of an MMCAI, and we found higher concentrations in 30-day non-surviving patients (Lorente et al., 2016). Therefore, the determination of serum TAC concentrations at fourth and eighth day of MMCAI is a new aspect in our study.

Furthermore, higher serum TAC concentrations also at days four and eight of severe MMCAI in non-surviving patients were novel findings in our study. Another interesting and new finding in our study for clinicians was that serum TAC concentrations in the first week of severe MMCAI could be used to predict mortality. These novel findings from our study may be due to our inclusion of patients only with an acute infarction $\ge$ 50% in the middle territory and with GCS $\le$ 9.

Another interesting and new finding in our study was the correlation between lipid peroxidation (estimated by serum concentrations of malondialdehyde) and serum TAC concentrations during the first week of severe MMCAI. A study has found higher plasma TAC concentrations and higher oxidative stress, assessed by plasma concentrations of nitric oxide metabolite levels (nitrite and nitrate), in diabetic acute stroke patients than in healthy control subjects (Guldiken et al., 2009). We propose like Guldiken et al. (2009) that the high serum TAC concentrations in ischemic stroke patients and non-surviving patients and its correlation with lipid peroxidation may be due to high reactive oxygen species production, and an attempt to avoid those harmful effects.

The blood-brain barrier (BBB) is a selective semipermeable border that separates peripheral blood and neural brain tissue in the central nervous system. The BBB restricts the passage of pathogens and large or hydrophilic molecules and the impermeability of the BBB has hampered the use of antioxidants agents for neuroprotection (Rashid et al., 2014). However, some murine model studies have shown that after the oral administration of tea, different substances of tea are brain permeable and have neuroprotective effects (by increasing antioxidant capacity in brain samples) (Pervin et al., 2019; Rashid et al., 2014). Oxidative stress and antioxidant capacity in brain samples have been reduced in cerebral ischemia animal models with melatonin administration (Blanco et al., 2017; Kryl’skii et al., 2019). Also, oxidative stress measured in plasma has been reduced by different antioxidant vitamins administered orally in patients with acute ischemic stroke (E, C, B2, B6, B12) (Ullegaddi et al., 2004, 2005, 2006). It is possible that the BBB damage that occurs in cerebral ischemia (and that could contribute to hemorrhagic transformation) (Li et al., 2019) could favor the passage of those substances through BBB and reduce oxidative stress in cerebral ischemia. We found high serum concentrations of TAC and malondialdehyde concentrations in non-survivor patients; however, we have not assessed antioxidant capacity and oxidative stress in brain samples in our study.

The 2007 American guidelines on ischaemic strokes (Adams et al., 2007) suggested the use of intravenous thrombolysis therapy in patients without multilobar infarction in CT (hypodensity in a third of cerebral hemisphere). The improved outcome with thrombectomy is unclear, and in severely affected patients due to malignant edema, decompressive surgery is recommended. The American guidelines of 2018 (Powers et al., 2018) suggested the use of intravenous thrombolysis therapy in patients with an ischemic injury involving less than one-third of the middle cerebral artery territory; the use of thrombectomy in patients with occlusion of the internal carotid artery or middle cerebral artery infarction segment 1, pre-stroke modified Rankin scale $\le$ 1, age $\ge$ 18 years, Alberta Stroke Program Early Computed Tomography Score $\ge$ 6, National Institutes of Health Stroke Scale $\ge$ 6, and treatment initiated within 6 hours of symptom onset; and the use of decompressive craniectomy in patients with $\le$ 60 years of age, unilateral middle cerebral artery infarction and neurological impairment within 48 hours despite medical therapy. In our series (anterior to the new recommendations of the 2018 American Guidelines) only MMCAI patients without thrombectomy treatment had no significant differences in the survival rate and this was found between patients with or without intravenous thrombolysis, and with or without decompressive craniectomy.

There are some limitations in our study. We have not recorded the time interval from the onset of stroke to blood draws for the TAC measurement, to surgical intervention, and to hemorrhagic transformation to evaluate whether those factors could modify serum TAC levels. We have not registered the number and causes of exclusion and the percentage of severe MMCAI about the total ischemic stroke patients. Patients with autoimmune diseases or sepsis were excluded from the study, and we did not find differences in renal or hepatic functions at inclusion between surviving and non-surviving patients; however, data on atrial fibrillation, coronary artery disease, smoking and dyslipidemia which could influence serum TAC levels were not provided.

We chose to include only patients with severe MMCAI because of the higher mortality in that group of patients (50% in our series). Serum TAC levels could have potential predictive capabilities within the population under study; however, the results of this study are not generalizable to all strokes due to exclusion criteria. Also, the results of this study need to be replicated by another team before using them in the prediction of mortality due to the scarce and contradictory data on blood TAC concentrations and prognosis of patients suffering from ischemic stroke. We think that the interest of this study is not only that serum TAC concentrations could potentially predict mortality, but also to explore the effect of new modalities of treatment in this subgroup of patients at high risk of mortatility.

The potential use of serum TAC concentrations at any time during the first 7 days of a severe MMCAI infarction without thrombectomy to predict mortality was the main novel finding of our study.

APACHE: Acute Physiology and Chronic Health Evaluation; aPTT: activated partial thromboplastin time; FIO${}_2$: fraction inspired of oxygen; GCS: Glasgow Coma Scale; ICU: Intensive Care Unit; INR: International normalized ratio; PaO${}_2$: pressure arterial of oxygen; TAC: total antioxidant capacity.

LL coordinated and designed the project, and drafted the article. JJC, JSV, MA, MMM, LRG, VGM, and RS participated in data recollection. PAG, APC, and AFGR performed blood determinations. AJ participated in data interpretation. Finally, the manuscript was critically revised and approved by all authors.

Our perspective and observational study were developed in the Intensive Care Unit of six Spanish hospitals after the approval of the Institutional Ethical Board of all hospitals and with the written informed consent of a family member of each patient.

Grants of Fundación DISA a la Investigación Médica 2017 (Tenerife, Spain) (code: OA18/011) and Instituto de Salud Carlos III (Madrid, Spain) (code: PI18-00500). The funders had no influence in the design, analyses or preparation of the article.

The authors have not competing interests.