Inflammation and rehabilitation outcomes in patients with nontraumatic intracranial haemorrhage

1 Introduction

Despite considerable advances in the diagnosis and treatment of hemorrhagic strokes, outcomes relative to death and disability remain poor (Mathers et al., 2000; American Heart Association, 1996). For physical disability, 59% of patients with subarachnoid haemorrhage (SAH) have a poor outcome at 3 months after the acute event (Badjatia et al., 2010). Early studies have found total disability in about 15% of patients with SAH, and moderate/good recovery in 20– 35% of patients with hemorrhagic brain-injury (HBI) (Rosenørn et al., 1987; Säveland et al., 1986). In survivors, the severity of both brain and clinical damage during acute HBI, are important causal factors of reduced physical autonomy. Indeed, the acute event primes intracranial and systemic inflammatory response syndrome (SIRS), (Vila et al., 2000; Leira et al., 2004; Silva et al., 2005) which adds to systemic and intracranial pathophysiological processes (i.e., elevated intracranial pressure, reduced cerebral blood flow, decreased tissue oxygen supply, decreased total systemic blood flow) (Daverat et al., 1991).

Similar to ischemic strokes, HBI inflammation predicts early neurological deterioration (Leira et al., 2004). Of particular importance, plasma pro-inflammatory cytokines (TNFα, IL6) in subjects with HBI are associated with subsequent hematoma enlargement (Silva et al., 2005). The problem is that inflammation is particularly prevalent in acute strokes. Indeed, it develops in 59% of SAH patients admitted to intensive care unit (ICU) (Yoshimoto et al., 2001) and may also develop in up to 85% of the patients during their first 4 days of ICU stay (Dhar & Diringer, 2008).

Today, it has not yet been documented whether a possible persistence of inflammation in subacute (>3 weeks after acute events) patients negatively influences long-term functional recovery. This is crucial for the rehabilitation outcome of HBI subjects, as inflammation can potentially be manipulated by metabolic (Aquilani et al., 2015), drug or neurophysiological (vagal stimulation) interventions. As a result, the objective of this retrospective study conducted on rehabilitation subjects with HBI, was to investigate both the prevalence of residual systemic inflammation and its impact on patient recovery of functional independence.

2 Patients and methods

3 Results

Forty-two individuals with HBI were screened for the study. Eleven were excluded because of their provenience from settings other than ICU. As a result, thirty-one subjects were eligible for the study and were analyzed. Nine subjects were on enteral nutrition (29%) and 22 (71%) had oral alimentation.

Table 1 reports the demographic-, anthropometric-, clinical-, biohumoral-, nutritional-, functional characteristics of the entire population, both at discharge from ICU/admission to Rehab and at discharge from the Rehab institute.

At discharge from ICU/admission to Rehab (Table 2), a large majority of patients (74.2%) showed a persistence of mild systemic inflammation as indicated by serum IL-6, CRP levels above normal serum concentrations, which was significantly higher (p <  0.05 both) than in non-inflamed (25.8%) patients. In addition, erithrocyte sedimentation rate (ESR), white blood cell count, neutrophil and monocyte counts, neutrophil/ lymphocyte ratio (N/L) were all significantly higher in inflamed patients. Inflammation was associated with increased amounts of circulating positive proteins of the acute phase response (α1 globulin, fibrinogen) and respectively, decreased levels of negative ones (albumin, prealbumin). The inflamed patients were mildly under-nourished (actual BW <95% of habitual BW), whereas non-inflamed subjects were not. Patients as an entire population had important reductions in their functional independence (FIM), (Table 1) but this reduction was more pronounced in inflamed subjects (p <  0.05) (Table 2).

Functional independence negatively correlated with serum α1-globulin (r = – 0.565, p = 0.0007), N/L ratio (r = – 0.52, p = 0.002), neutrophils (r = – 0.47, p = 0.01), serum CRP (r = – 0.365, p = 0.01) and positively linked to albumin (r = 0.43, p = 0.008). At regression analysis, α1-globulin (p <  0.01), N/L ratio (p = 0.05) and albumin (p = 0.02) remained significantly correlated with FIM.

At discharge from Rehab, (Table 2), all inflammation indicators in inflamed subjects decreased, although CRP and IL-6 did not normalize. Even given these molecules were not different from non-inflamed subjects, they did indicate the permanence of residual inflammation. Coherently, circulating proteins of acute phase response improved with reductions of positive reactants and the increase of negative ones. In more detail, the fibrinogen concentration of inflamed group, compared to that in non-inflamed one significantly decreased over time. In the inflamed group, the blood glucose significantly changed.

The patients as an entire population, (see Table 1) improved their functional independence by about a score of 25.9. The gains in FIM (ΔFIM) were similar for both inflamed (+28.1) and non-inflamed (+19.5) subjects (p = 0.2). The results moreover, showed that 54.8% of the HBI patients did not reach the minimum gain of +22 score FIM (Beninato et al., 2006), which was considered clinically significant (Aquilani et al., 2014b). In these patients, the ΔFIM was 9.35±6.58 scores, quite unlike the ΔFIM of subjects (45.2%) with ΔFIM ≥22 scores (+45.93±19.17 scores; p <  0.0001). During rehabilitation, improvements in functional independence of the patient population were linked to improvements of circulating albumin only (ΔFIM vs Δalbumin r = +0.56, p = 0.0005).

The study showed that at discharge from ICU, the prevalence of residual inflammation among subjects with HBI can reach significant levels, thus contributing to patient disability. However, during the rehabilitation stage of the disease, residual inflammation does not seem to hinder patient recovery of functional independence given that both, inflamed and non-inflamed subjects had similar opportunity to retrieve their physical capacity rate.

1) Inflammation and functional independence

a) At discharge from ICU/admission to Rehab

Disease severity (Badjatia et al., 2010), neurological interventions (Esper et al., 2006) and infection complications (Esper et al., 2006; Iadecola & Anrather, 2011) during ICU-stay are probably the major factors responsible for the persistence of inflammation at patient discharge from ICU. At this stage of the disease, inflammation was mild and accounted for alterations of circulating proteins, both positive and negative, of the acute phase response. The fact that α1-globulin and peripheral inflammatory neutrophils and monocytes where high, might suggest the persistence of inflammatory processes in the damaged brain area at about 30 days after the acute event. This agrees with pathological studies showing an inflammatory infiltrate in the lesioned brain area that persists for years after any stroke (Iadecola & Anrather, 2011). On the other hand, a high N/L ratio indicates a prevalence of inflammatory activities over the adaptive immunity and repair processes (el-Hag & Clark, 1987). This is reinforced by the finding of increased α1-globulin concentrations, which indicates increased levels of the anti-protease system, which in turn limits protease overproduction/ activity of neutrophils and monocytes primed by the inflamed brain area (Iadecola & Anrather, 2011). Inflammation probably impacts the development of patient disability, as the FIM score was lower in inflamed compared to non-inflamed ones, and various indicators of inflammation negatively correlated with FIM. The relationship between α1-globulin and functional independence was also found in subjects with cerebral ischemia (Aquilani et al., 2014a).

b) At discharge from Rehab

During rehabilitation, the inflammation rate declined but protein changes of the acute phase response remained, even if they were attenuated. Reduced inflammation was associated with functional independence improvement, similar to non-inflamed subjects. In inflamed subjects, the reduced inflammation created a better blood glucose concentration, suggesting improved cell glucose metabolism.

This study cannot explain the similar gains of functional independence between inflamed and non-inflamed subjects. A potential explanation could be the ambivalent activity of inflammation, which on one hand is detrimental to the injured brain although at the same time, may also favour brain repair (Aquilani et al., 2014a). Indeed, inflammatory cytokines have concomitant anti-inflammatory properties as they down-regulate inflammatory tumour necrosis factor-α (TNF-α) and IL-1 (Zoico & Roubenoff, 2002) and induce the production of anti-inflammatory glucocorticoids (Chang & Bistrian, 1998). The pro-inflammatory IL-6, increased in our inflamed patients, exerts direct neuroprotection activities and regulates learning and memory process by influencing both nerve cell regeneration after injury (Chang & Bistrian, 1998) and neuronal differentiation (Kishimoto, 1989), increasing secretion of neurotrophic factor by astrocytes (Kishimoto, 1989). Indeed, the IL-6 is highly expressed in the hippocampus (Bliss & Collingridge, 1993) cerebral cortex (Kirkwood et al., 1995), amygdale (Stevens, 1998) and all structure involved in cognitive processes.

Thus, the mild increase of IL-6, observed here, may have played a role in inducing neuroprotection explaining in this way the improvement rate of FIM in inflamed subjects. On the other hand, normalized circulating phagocytic cells and N/L ratio, associated with improved FIM, would suggests that during the rehabilitation stage of HBI, the repair processes in the brain’s damaged area probably prevailed over inflammatory ones (adaptive brain remodelling).

Trends towards normalization of liver synthesis of proteins of acute phase response, is of great clinical importance, given that any gain of functional independence was linked to the improvement the supply of albumin. Therefore, patient recovery of functional capacities appears to depend upon, both the reduction of inflammation, leading to a prevalence of adaptive immunity (repair processes), and on increases in circulating albumin. Interestingly, a correlation between overtime changes in albumin and in functional independence was also found in individuals with sub-acute cerebral ischemia (Aquilani et al., 2014a).

2) Circulating albumin levels and functional recovery

The albumin contribution to improving post-HBI recovery of brain functioning should not be underestimated for at least two reasons. Firstly, hypo-albuminemia may be caused not only by inflammation but also due to a patient’s inadequate protein-calorie intake, a condition which can easily be manipulated. Secondly, albumin exerts important neuroprotection activity.

Inadequate protein-calorie intake frequently occurs in both acute (Davalos et al., 1996; Davis et al., 2004) and subacute (Aquilani et al., 1999, 2008a) stroke and conditions the clinical (Davalos et al., 1996; Davis et al., 2004) and functional prognoses of patients (Aquilani et al., 2008a, b). Liver albumin synthesis may also be impaired in cases of alterations in plasma amino acid profile (Aquilani et al., 2014b). It has been documented that adequate protein/ amino acid intake exerts direct action on neuronal survival (Doutheil et al., 1999) and neuro-cognitive recovery of subjects with ischemic stroke (Aquilani et al., 2008a, 2010). The mild undernutrition state of our inflamed patients, showed at admission to, and at discharge from Rehab, would suggest that patient protein-energy intake was not optimal and could limit the improvement of liver albumin synthesis due to reduced inflammation. Moreover, inadequate protein energy intake may also have contributed to lower serum prealbumin.

The important direct neuroprotection exerted by albumin has been established in several studies showing that human albumin therapy provides neuroprotection in transient focal ischemia (Belayev et al., 1997, 1998, 2001) global ischemia (Belayev et al., 1999) and traumatic brain injury (Ginsberg et al., 2001). As a result, sub-optimal protein energy intake and therefore, lack of albumin improvement could limit brain repair, which could in turn contribute to explain why most patients did not achieve the minimum FIM gain (≥22 scores) translating into clinical advantage to patients (Beninato et al., 2006), and why mean FIM gain in patients with increased FIM ≥22 scores was only 45. We speculate that intensive rehabilitation therapy, when not accompanied with optimal nutritional metabolic conditions, may not be enough to achieve better rehabilitation outcomes. This could explain the loss of physical autonomy of our study patients. The disability rate was 60% in subjects with ΔFIM <22, and 37.5% in those with ΔFIM ≥22. This study would suggest that optimized protein/amino acid intake, timely improvements of circulating albumin and modulation of neuro-inflammation (el-Hag & Clark, 1987) could all be useful tools to enhance rehabilitation outcomes. These issues, doubtlessly require further well-planned investigation. An additional suggestion from this study is that these interventions should be initiated from the moment the patients are accepted into ICU.

5 Conclusions

This study shows that the prevalence of a residual inflammation was shown to be high in subjects with non-traumatic cerebral haemorrhage at discharge from ICUs, contributing to the patient disability rate. Its reduction during rehabilitation, would favour patient functional capacity recovery.

Conflict of interest

All authors declare no conflicts of interest.