All procedures were approved by the local review board (Comité Ético de Investigación Clínica de La Rioja, Logroño, Spain) and all described procedures adhere to the tenets of the Declaration of Helsinki.

The present study was designed as a prospective, observational, and longitudinal clinical study with patients diagnosed with acute ischemic stroke at the Neurology Service of the Hospital San Pedro (Logroño, Spain) from October 2014 to April 2015. Consecutive patients (n=76) fulfilling inclusion criteria signed the written informed consent documents and were recruited into the current study. Inclusion criteria required patients to be suffering from acute ischemic stroke demonstrated by magnetic resonance imaging (MRI), and an evolution of <24 h. Exclusion criteria were the same used for the hospital's stroke unit and were as follows: Contraindications for performing MRI; age, <18 years; dementia; previous stroke within 3 months; cranial traumatism; central nervous system infection; cardiac insufficiency; renal failure; sepsis; active neoplasia; active inflammatory or autoimmune disease; pregnant or lactating women; and patients who would not be able to complete proper follow-up.

In this time period, 140 patients were diagnosed and treated at the stroke unit. Of these, 127 were suspected of suffering ischemic stroke and 120 arrived within 24 h of the first symptoms. A total of 29 patients were excluded from the present study due to fulfilling at least one of the exclusion criteria and 15 additional patients were excluded following performing MRI as the ischemic lesion could not be confirmed. Finally, 76 patients were included in the study and all required data was collected for 70 of them.

Patients received standard care following the approved protocols of the stroke unit. General characteristics of the patient were collected as part of the clinical history (including age, gender, risk factors, current medical treatment and previous functional situation). During their stay at the stroke unit, numerous parameters were continuously monitored, including electrocardiogram, systolic arterial pressure (SAP) and diastolic arterial pressure (DAP), temperature, and hypoxemia. Neurological severity was measured with the National Institutes of Health Stroke Scale (NIHSS) (22) at 0, 24 h, 7 days, and 3 months. Functional prognosis was also evaluated with the Rankin scale at 3 months (23). In addition, blood plasma samples were taken on day 1, day 2, and day 7 to quantify circulating levels of NO and AM. Infarct volume evolution was established by comparing the images captured by MRI on day 1 and at day 7 with a 3 Tesla instrument (Discovery MR 750w; GE Healthcare Life Sciences, Chalfont, UK).

NO production was indirectly quantified by determining NOx, using an ozone chemiluminescence-based assay adapted to plasma samples (24). Briefly, plasma samples were deproteinized with 0.8 N NaOH and 16% ZnSO₄ solutions (1/0.5/0.5, v/v/v; Sigma-Aldrich, Madrid, Spain). Following centrifugation at 10,000 × g for 5 min at room temperature, the resulting supernatants were removed for chemiluminescence analysis in a NO analyzer (NOA™ 280i; GE Analytical Instruments, Boulder, CO). NOx concentration was calculated by comparison with standard solutions of sodium nitrate. Final NOx values were expressed in µM.

The concentration of AM present in blood plasma was determined using a commercially available radioimmunoassay (RIA) kit (Phoenix Europe GmbH, Karlsruhe, Germany). Samples (1 ml) were initially diluted in an equal volume of 0.1% alkali-treated casein in phosphate-buffered saline at pH 7.4, and applied to pre-washed reverse-phase Sep-Pak C-18 cartridges (Waters Corporation, Milford, MA, USA) to remove the AM-binding protein, complement factor H. The peptide fraction was eluted from the C-18 matrix with 3 ml 80% isopropanol containing 0.125 N HCl and freeze-dried overnight, as previously described (25). AM levels contained in lyophilized extracts were then determined by RIA following the manufacturer's protocols.

All statistical analyses were performed using the SPSS version 21 software package (IBM SPSS, Armonk, NY, USA). A descriptive analysis of all variables was performed and categorical variables were expressed as absolute and relative frequencies. Continuous variables were defined by their mean and standard deviation (SD) when their distribution was normal (as tested by the Shapiro-Wilk test) or as the median and interquartile range when the distribution was not normal. Univariate analyses were performed with Pearson's χ² test, modified by Fisher's exact test, or in the case of continuous variables with Student's t-test or analysis of variance. When samples did not follow a normal distribution, non-parametric tests, including Kruskal Wallis followed by Mann-Whitney's U test were performed. Parameters that were marked as significant in univariate analysis entered multivariate binary logistic regression to investigate independent effects. A receiving operating characteristic (ROC) analysis was performed to assess the potential predictive value of binary classifier systems. P<0.05 was considered to indicate a statistically significant difference.

The clinical sample included 76 stroke patients, 33 women (43.4%) and 43 men (56.6%), with a mean age of 73.3±12.3 years (Table I). Age-matched healthy controls were 5 women (35.7%) and 9 men (64.3%) with a mean age of 73.4±12.2 years. Certain patients had been exposed to relevant risk factors, including arterial hypertension, dyslipidemia, diabetes, atrial fibrillation, ischemic cardiopathy, or a previous stroke (Table I). A marked proportion of these patients had been treated with antiaggregants, anticoagulants, antihypertensives, and statins (Table I). Following completion of the etiological profile, ~1/3 of the patients were diagnosed with atherothrombotic stroke, another third with cardioembolic stroke, and 14.5% with lacunar stroke (Table I). The median NIHSS taken at admission was 3, with a median infarct volume of 3.8 cm³ (Table I). Intracranial or extracranial occlusion of a large vessel was demonstrated in 17 patients (22.3%).

At admission in the emergency room (ER), mean values for SAP and DAP were 160.1±30.0 mmHg and 84.8±18.7 mmHg, respectively (Table I). Hypertension was observed at the admission of 59 (77.6%) patients. Women exhibited significantly higher SAP values than men at admission (P=0.04), and on day 1 (P=0.008) and day 2 (P=0.003) of treatment; however, this difference was corrected by day 3. No significant differences by gender were observed for DAP (Fig. 1A). Patients with a history of hypertension exhibited significantly higher SAP values, as compared with normotensive patients at admission (P=0.03), and on days 1 (P=0.004) and 2 (P=0.002) of treatment (Fig. 1B). DAP values were similar independently of hypertension history. There was a significant reduction in SAP and DAP (P<0.001), during the first day of treatment, and a further reduction in SAP (P<0.05) from day 1 to day 2 (Fig. 1C). When comparing blood pressure and neurological severity measured by the NIHSS at 24 h, patients with more severe symptoms (NIHSS >3) had significantly higher SAP (Fig. 1D; ER, P=0.002; day 1, P=0.014; and day 2, P=0.005) and DAP (data not shown) until day 2, with this trend reversing at day 3 (Fig. 1D; P=0.009), as compared with patients with less severe symptoms (NIHSS ≤3).

Variability in blood pressure (BP) values has been proposed as a possible cause of brain deterioration in ischemic stroke patients (26). This variability is established by measuring the SD, the successive variation, and the coefficient of variation of either the SAP or the DAP. Notably, almost all variability parameters were significantly higher in women when compared with men (Table II). This higher variability was also observed in patients with increased neurological severity as measured by the NIHSS at 24 h and at 7 days (Table II). In addition, patients whose infarct volume grew during the first 7 days had significantly higher BP variability than those whose infarct volume did not grow (Table II). There was no association between the variability in BP during the stay of the patients in the stroke unit and their performance at 3 months, measured by either the NIHSS or the Rankin scale (data not shown).

In healthy control subjects, NOx levels were 15.3 µM (12.1–17.9). These levels were significantly higher in women than in men (P=0.02; data not shown). In stroke patients, NOx levels measured on day 1 and day 2 of their hospital stay were significantly lower (P=0.008) than those obtained in healthy subjects. At day 7, NOx levels were significantly higher than at day 1 (P<0.001) and day 2 (P<0.01), but indistinguishable from healthy controls (Fig. 2A). In stroke patients, no differences were observed with gender.

When NOx levels were compared with neurological severity, it was observed that patients with an NIHSS ≤3 at day 1 had significantly higher values of NOx on day 2 (Fig. 2B; P=0.02). There was also an association with BP parameters. Patients in the third tertile of NOx levels at day 1 had a significantly lower DAP on days 2 (P=0.009) and 3 (P=0.02; Fig. 2C). Patients in the third tertile of SAP at admission had significantly lower (P=0.01) NOx levels than other patients in the second and first tertile (Fig. 2D).

Healthy control volunteers had a median AM value of 389.7 pg/ml (343.9–475.9). No differences were observed for either gender or age groups. In stroke patients, AM levels were significantly higher than in healthy controls at day 1 (P=0.003) and 2 (P=0.005), however, these values were indistinguishable from controls at day 7 (Fig. 3A). Among patients, women had significantly higher AM levels than men (P=0.02; data not shown). Notably, patients that had undergone treatment with antiaggregants or statins previous to the stroke had significantly lower levels of AM as measured at day 2 (P=0.04 and P=0.001, respectively; data not shown). Patients with a history of hypertension had similar levels of AM to the other stroke patients, however, their AM concentration decline from day 1 to 7 was significantly steeper than in other patients (P=0.005; data not shown).

There was no association between AM levels and neurological severity as measured using the NIHSS at admission. However, patients with a high NIHSS score (>3) at day 1 had significantly higher AM levels at day 2 (Fig. 3B; P=0.014). A similar association was observed with NIHSS values measured at day 7 and AM values obtained on day 1 and day 2 of treatment (Fig. 3C; P=0.012 and P=0.002, respectively). These data suggested that AM levels measured on day 2 may predict neurological severity at day 7. To investigate this hypothesis, a ROC curve analysis was performed and it was demonstrated that the optimal threshold for AM levels was 522.13 pg/ml. This value renders an area under the curve of 0.721 (95% confidence interval, 0.590–0.852), a sensitivity of 70%, and a specificity of 60% (Fig. 3D).

Infarct volume growth was defined as the difference between the infarct volume measured by MRI at day 7 and the initial volume measured on day 1 of treatment. Patients were classified as having infarct volume growth when they exhibited a ≥20% increase, 26 patients (37.1%) experienced such growth while 44 (62.9%) did not. Demographic, clinical, and analytical characteristics of the patients were examined to investigate whether any had predictive value over the potential infarct volume growth as measured at day 7 (Table III). Notably, among the reported risk factors, only a previous treatment with antihypertensives or antiaggregants resulted in a clear protection against infarct volume growth (P=0.04 and P=0.01, respectively). There was a linear association between NIHSS scores (>3), either at admission or at 24 h, and volume growth. There was also an association with elevated diastolic blood pressure but not with the systolic component. In addition, patients with occlusion of a large vessel were more likely to experience infarct growth (P=0.02). The association with NOx was notable as there was no association with specific NOx levels but with the increment in NOx concentration from day 1 to 7 (P=0.03). No association was observed with AM levels.

Subsequent to conducting univariate analysis, multivariate logistic regressions were performed to identify independent predictors of clinical outcomes. These outcomes were the growth of the infarct volume (day 7 - day 1), the NIHSS scores measured at day 7 and at 3 months after the onset of the condition, and the Rankin scale, also measured at 3 months (Table IV). As predictors of infarct volume growth it was demonstrated that pretreatment with antiaggregant therapeutic agents protects patients, whereas a high NIHSS score taken at 24 h, the confirmation of the occlusion of a large vessel, and the steep increase on NOx levels provide a poorer prognosis. For neurological severity measured at day 7, it was observed that male gender and the increase in NOx levels from day 1 to day 2 are protective, whereas a high NIHSS score taken at 24 h, confirmation of the occlusion of a large vessel, and the increase of AM levels from day 1 to day 2 are markers of a poorer prognosis. For the neurological severity measured at 3 months the increase in NOx levels from day 1 to day 2 is a positive marker for improved recovery, whereas a high NIHSS score at 24 h and increased AM levels are negative predictors. As measured by functional prognosis at 3 months, male gender and a high increase in NOx levels from day 1 to day 2 are reliable predictors for a good prognosis (Table IV).

It was observed in the present study that patients that had been treated with either antihypertensive or platelet antiaggregant therapeutic agents prior to stroke onset were protected against the growth of the infarct volume. It has been demonstrated that in-hospital treatment with these therapeutic agents may explain the decrease in mortality from ischemic stroke (27), and that taking these therapeutic agents prior to stroke onset may have a marked positive prognosis (28,29). Data from the present study confirms these previous observations.

In the patients investigated in the present study, women had a higher SAP than men at admission. This gender bias has been widely reported, despite men and women having similar risk factors (30). This difference between genders and the commonly high values in BP were progressively controlled during day 1 and 2 of stay in the stroke unit. The success of stroke units is a consequence of the control of all physiological factors that may influence the evolution of the ischemic lesion, referred to as non-pharmacological neuroprotection (31), which includes monitoring BP. Patients with higher BP presented higher values of neurological severity measured at 24 h, suggesting that lowering BP may be a good strategy to reduce brain injury. However, the scientific literature is divided on this issue. The majority of observational studies associate high BP with poor clinical outcomes (32,33). By contrast, other reports show improved clinical evolution with higher BP levels (34), and low values of BP have been also associated with a poor prognosis and higher mortality (35). A previous study has suggested that a U-shaped association may be present, with high BP values inducing brain edema and hemorrhagic transformation and low BP values contributing to the transformation of the penumbra into infarcted area (36). Previous interventional studies using antihypertensive therapeutic agents suggest that reducing BP during the acute phase is safe and may reduce mortality at 3 months (37,38), whereas other previous studies have demonstrated either no benefit (39) or a small increase in early adverse events (40). A recent review, which combines data from 26 articles and 17,011 patients, concluded that there is no evidence that reducing BP during the acute phase of the stroke may save lives or reduce disability (41).

However, the variability of DAP and SAP values during the acute phase may provide more information. In the present study, an increased BP variability was observed in women, in patients with higher NIHSS scores, and in patients where infarct volume growth was reported. This is in agreement with previous reports where BP variability has been associated with poor clinical outcome (26,42,43).

The present study has demonstrated that peripheral concentrations of NOx in stroke patients are lower than in the healthy control population. Previous studies disagree on whether peripheral NOx increases or decreases following stroke. Certain previous studies agree with the results from the present study indicating lower NOx levels in patients (44,45), whereas other studies have described elevated levels compared with healthy controls (46,47). This discrepancy may be due to different methods of measuring NOx or to variations in the L-arginine pathway, as a result of endothelial dysfunction, leading to elevated levels of symmetric and asymmetric dimethylarginine. These metabolites are elevated in stroke patients and result in reduced production of NO (48). Depending on the quantity of these metabolites, the final levels of NOx in the blood may vary. In addition, the potential contributions of NOS-independent sources of NO must be considered (49). In the patients investigated in the current study, during the first week of the follow-up period, NOx levels returned to normal, suggesting a progressive recovery of homeostasis. This elevation of NOx levels post-stroke is in agreement with previous studies (14).

The present study demonstrated that elevations of NOx levels may be beneficial or detrimental for the patient, depending on when they occur. Elevation of NOx from day 1 to day 2 was demonstrated to be protective and it predicted a positive outcome at 7 days and 3 months, for neurological recovery and functional dexterity. However, an elevation of NOx from day 2 to day 7 predicted growth of the infarct volume. This dual behavior may be associated with the type of NOS isoform that is activated. The initial elevation of NOx production may be associated with the endothelial NOS, an isoform that has been demonstrated to be neuroprotective via exerting an effect on vasodilatation, inhibition of platelet aggregation, and induction of angiogenesis (50). The elevation occurring from day 2 to 7 may be more associated with the activation of the inducible NOS isoform, the activity of which is initiated at ~12 h after ischemic onset and continues for up to 8 days later (51). This isoform is considered damaging to the surrounding tissue due to the unregulated large quantities of NO produced (7).

As expected, an inverse association between NOx levels and BP was observed, confirming the vasodilatatory effect of NO in a clinical setting, as has been thoroughly reported in experimental models (7,9).

The present study observed that stroke patients had significantly higher values of plasma AM than healthy individuals, this is in agreement with previous literature (20,52). By day 7, AM levels returned to normal, suggesting a progressive resolution of the pathology. A high AM level at day 2 or an increase in AM levels from day 1 to day 2 was associated with increased neurological severity at day 7 and at 3 months after the stroke. This was confirmed with a ROC curve analysis, where high AM values predicted a poorer outcome. A previous study has demonstrated that the AM level in the blood of ischemic stroke patients may be also predictive of 3-month mortality and unfavorable outcomes (52), suggesting AM is a negative predictor of clinical performance.

In experimental animals, it has been identified that eliminating AM from the central nervous system results in larger infarcts following acute ischemic stroke (21) and that injecting AM reduces infarct volume (53). These studies indicated the neuroprotective effects of AM are due to its vasodilatatory, pro-angiogenic, and anti-apoptotic effects. However, another previous study indicated intracerebroventricular injection of AM resulted in an increase of ischemic injury (54). This detrimental effect of AM may be mediated by its modulatory action on the immune system (55), which is important for stroke resolution (56). Overexpression of AM has been detected in leukocytes of stroke patients and this higher expression is associated with stroke severity (57). The clinical observations suggest that, during human stroke, AM levels may be a consequence of stroke severity, despite the neuroprotective effects of AM on the infarcted tissue.

The current study observed that patients taking antiaggregants or statins prior to stroke onset had lower levels of AM measured at day 2. Statins are well known for reducing cholesterol levels, however, they also exert pleiotropic actions, including antioxidative and cellular protective effects. These characteristics suggest that statins may provide novel therapeutic approaches for various neurological disorders, including stroke (58). There is little information regarding the association between AM and statins. Using AM heterozygous knockout mice, Yamamoto et al (59) demonstrated that statins significantly inhibited fibrosis and apoptosis while inducing angiogenesis in a model of heart fibrosis. The underlying mechanism for the protective action of statins in stroke may include a reduction of AM levels.

In conclusion, BP variability and temporal profiles of NOx and AM levels have been demonstrated as predictors of clinical outcomes in stroke patients, as measured by infarct volume growth, neurological NIHSS scales at 7 days and 3 months, and functional prognosis at 3 months. Development of rapid tests for evaluating NO and AM levels may be useful for predicting patient outcome, for developing personalized therapeutic strategies, and for stratifying stroke patients in clinical trials.