Introduction

Subarachnoid hemorrhage (SAH) is a prominent cause of morbidity and mortality worldwide, with overall incidence reported to be about 9 per 100,000 person-years [1] and mortality estimated to be about 35% [2]. In about 85% of cases, the underlying cause is ruptured intracranial aneurysm, which is associated with poorer outcomes than non-aneurysmal cases (perimesencephalic and various other types) [3]. Development of aSAH is shown to peak in middle-aged and older adults, particularly between the ages of 50 and 60 years, and is 1–6 times more common among older adult women than men [1].

Although clinical decline after aSAH may arise from etiologies such as hydrocephalus, rebleeding, metabolic disturbances or infection [4], delayed cerebral ischemia (DCI) is a frequent cause of poor neurological outcomes and death in patients who survive the initial bleeding of aSAH [5]. DCI also may progress to cerebral infarction, resulting in severe disability or death in about 50% of SAH patients, and is deemed the most important adverse prognostic factor for poor outcome of aSAH [6]. Patients who develop DCI, may manifest unifocal or multifocal areas of ischemia, typically between days 4 and 14 after the initial hemorrhage, and not restricted to the arterial supply or borderzone areas [4]. Although vasospasm figures prominently in DCI development, the pathogenesis is believed to be multifactorial [7, 8].

The currently accepted predictive factors for long-term neurologic outcomes after SAH, especially DCI and poor functional outcomes, include age and clinical severity [9, 10]. A review study investigating hemostatic or fibrinolytic parameters that may predict DCI also reported that levels of von Willebrand factor and platelet activating factor were higher in SAH patients with DCI than in those without DCI [4]. Recent evidence of serum parameters that serve as biomarkers for predicting SAH outcomes suggests that various inflammatory, fibrinolytic and hematologic factors may be associated with poor outcomes after SAH, including elevated glucose level [11], low platelet count [12], elevated CRP [13,14,15], and white blood cell (WBC) count [15, 16].

Having a biomarker which is convenient to obtain in assessing patients’ individual risk for poor outcome after aSAH would have tremendous clinical value. Among the potential biomarkers, CRP and WBC are markers of systemic inflammation which are routinely collected on admission have drawn much attention of research. However, the relationships between these parameters at early phase of admission and the occurrence of DCI and poor outcome at follow-up remains ambiguous, and no meta-analysis has yet estimated them. To investigate whether these parameters may aid in identifying patients at high risk of poor prognosis, and to ultimately improve patient monitoring and individualize care, we conducted this systematic review and meta-analysis to determine the relationships between peripheral CRP, WBC, and the occurrence of DCI and poor functional outcome following aSAH.

Materials and methods

The present systematic review and meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. Pubmed, EMBASE, and CENTRAL databases were searched until November 30, 2019 using the following keywords and word combinations to maximize search results:

(((((C-reactive protein [MeSH Terms]) OR white blood cell [MeSH Terms]) OR ((biomarkers [MeSH Terms]) OR predictor)) OR inflammation [MeSH Terms])) AND subarachnoid hemorrhage [MeSH Terms]

The search filters applied were: Abstract available; Humans. Reference lists of relevant studies were also hand-searched for additional eligible studies.

Selection criteria and data extraction

Prospective and retrospective studies were selected that met the following inclusion criteria: (1) adults age > 18 years with spontaneous SAH due to ruptured aneurysm, diagnosed by computed tomography angiogram (CTA) or digital subtraction angiography (DSA); (2) reporting adjusted odds ratios (controlling for baseline characteristics such as age, sex, clinical grades of severity at admission, and/or other known predictors) for the associations between the occurrence of delayed cerebral ischemia (DCI) or poor functional outcome and blood CRP/WBC collected either on admission or at the early phase of admission as part of the routine clinical workup. Studies of patients with traumatic SAH or SAH due to other secondary causes were excluded; those focused only on results of cerebrospinal fluid (CSF) CRP or WBC and did not provide data of peripheral CRP/WBC, non-human studies, letters, commentaries, reviews, editorials, case reports, proceedings and personal communications; and studies with no quantitative outcomes of interest were excluded. Studies were identified by the above search strategy by two independent reviewers (XM and FL). Where uncertainty existed regarding eligibility, a third reviewer (YZ) was consulted.

Main outcome measures

Outcome measures included the occurrence of DCI and poor functional outcome. DCI was defined as a new onset of focal neurological deficit or a deterioration of consciousness (delayed ischemic neurologic deficit, DIND); and/or a new infarct on CT or magnetic resonance (MR) imaging that was not visible initially and was attributed to cerebral vasospasm (delayed cerebral infraction). Functional outcomes were graded by modified Rankin Scale (mRS) or Glasgow outcomes scale (GOS). Poor functional outcomes were defined as mRS 3 or 4 to 6, or GOS 1–3 from discharge to 1-year follow-up.

The following data were extracted from each included study: author, year of publication, study design, patient number, sex and mean age of subjects, time from ictus to admission, time from ictus to sampling, procedures performed, clinical severity (i.e., Hunt and Hess scale; WFNS scale) and radiological severity (i.e., Fisher grade) on admission, criteria used for poor outcome, exclusion criteria, time point of outcome evaluation, and factors adjusted in the associations analyzed. Of note, if multiple analyses on CRP/WBC collected at different time point were reported in the same study, the earliest timeframe was selected for the meta-analysis (Fig. 1).

Fig. 1
figure 1

PRISMA flow diagram

Full size image

Quality assessment

We assessed the quality of included studies using the Newcastle–Ottawa scale (NOS) for cohort studies as recommended by the Cochrane Non-Randomized Studies Methods Working Group [17]. This scale awards a maximum of nine stars to each study: four stars for the adequate selection of cohort participants, two stars for comparability of cohort participants on the basis of the design and analysis, and three stars for the adequate ascertainment of outcomes. Quality assessment was performed by two independent reviewers (XM and FL), and a third reviewer (YZ) was consulted if any uncertainties occurred.

Statistical analysis

To investigate associations of early increase in CRP and WBC with the occurrence of DCI and poor functional outcome following aSAH, odds ratios (OR) with 95% confidence intervals (CI) were extracted from the included studies. An χ2-based test of homogeneity was performed using Cochran’s Q statistic and I2. I2 illustrates the percentage of total variability in effect estimates among trials that is due to heterogeneity rather than chance. A rough guide to interpretation of I2 in the context of meta-analyses is as follows: (1) 0–40%: may not be important; (2) 30–60%: may represent moderate heterogeneity; (3) 50–90%: may represent substantial heterogeneity; (4) 75–100%: considerable heterogeneity [18]. In addition, a random-effects model was applied. Pooled effects were calculated and a two-sided P value < 0.05 was established as statistical significance. Sensitivity analysis was carried out using the leave-one-out approach. Publication bias analysis was not evaluated, because the number of studies was less than ten studies, too few to detect an asymmetric funnel [19]. All analyses were performed using Comprehensive Meta-Analysis statistical software, version 2.0 (Biostat, Englewood, NJ, USA).

Results

Among 42 eligible studies that underwent full-text review, 27 were excluded and 15 (5 retrospective, 10 prospective) that met the eligibility criteria were included [13,14,15,16, 20,21,22,23,24,25,26,27,28,29,30]. Characteristics of the included studies are summarized in Table 1. Overall, the 15 eligible studies analyzed the data of 3268 patients with aSAH. Patients’ median age was 53 years (range 48–57 years) and the percentage of older adult males ranged from 24 to 56%. Time from ictus to admission was within 24–96 h. Procedures performed, clinical and radiological severity are also summarized in Table 1. Parameters studied, timeframe on sampling, outcome definitions, exclusion criteria and factors adjusted among the included studies are summarized in Table 2. Among the studies included, 11 studies assessed the parameters of interest by initial blood sampling at presentation. 2 studies were based on blood samples collected on postoperative days 1–2, and 2 studies used blood samples collected during the first 14 days of admission. Clinical grade at presentation and patients’ age were the most frequently adjusted factors (for clinical grade: 11 studies; for age: 8 studies) (Tables 1, 2).

Table 1 Characteristics of studies included in this review and meta-analysis
Full size table
Table 2 Parameters studied, timeframe on sampling, outcome definitions, and factors adjusted
Full size table

Meta-analyses

Delayed cerebral ischemia

Five studies [13, 16, 20, 23, 24] provided adjusted ORs and were included in the meta-analysis to determine associations between early increase in CRP and the occurrence of DCI following aSAH. Substantial heterogeneity was found across the five studies (Q statistic = 11.428, I2 = 64.997%, P = 0.022). Early increase in CRP was significantly associated with higher risk of the occurrence of DCI (pooled OR 1.30, 95% CI 1.10–1.54; P = 0.002) (Fig. 2a).

Fig. 2
figure 2

Meta-analysis on the associations between occurrence of DCI and a CRP level and b WBC

Full size image

Only four studies [16, 24, 26, 27] provided enough data of peripheral WBC and were included in the meta-analysis to determine the associations between early elevated WBC and occurrence of DCI. Substantial heterogeneity was found among the four studies (Q statistic = 13.691, I2 = 78.089%, P = 0.003). Meta-analysis revealed there was no significant associations between increase in WBC and occurrence of DCI (pooled OR 1.13, 95% CI 0.95–1.34; P = 0.179) (Fig. 2b).

Poor functional outcomes

The criteria used for defining poor functional outcomes are summarized in Table 2. Nine out of 15 studies were included in the meta-analysis to determine associations between increase in peripheral CRP and poor functional outcomes after aSAH. Early increase in CRP was not significantly associated with poor functional outcome (pooled OR 1.02, 95% CI 1.00–1.04, P = 0.052; heterogeneity: Q statistic = 24.571, I2 = 71.511%, P = 0.001) (Fig. 3a).

Fig. 3
figure 3

Meta-analysis on the associations between poor function outcome and a CRP level and b WBC

Full size image

Results of meta-analysis of seven studies showed a significant association between early increase in WBC and increased risk of poor outcome (pooled OR 1.17, 95% CI 1.07–1.28, P = 0.001) with moderate heterogeneity among studies (heterogeneity: Q statistic = 13.864, I2 = 56.724%, P = 0.031) (Fig. 3b).

Sensitivity analysis and publication bias

Sensitivity analyses were performed using the leave-one-out approach in which the meta-analysis is performed with each study removed in turn (Table 3). The direction of combined estimates did not vary markedly with the removal of the studies, indicating that the meta-analysis was robust and the data were not overly influenced by any single study, only except for Sokół et al. [15] and Turner et al. [22] for CRP level associated with poor outcomes. After removing Sokół et al. [15] and Turner et al. [22], the results changed to significant.

Table 3 Sensitivity analysis
Full size table

Because of the small number of studies, the power of the tests for publication bias were too low to distinguish chance from real asymmetry.

Quality assessment

The result of quality assessment of included studies is shown in Supplementary Table 1. The total score ranged from 6 to 9 and the average score was 8, suggesting that the studies included were of good quality (Supplementary Table 1).

Discussion

In the present review and meta-analysis, we assessed the associations between DCI/poor functional outcome following aSAH and blood CRP/WBC. These two parameters were selected for meta-analysis, because they were the most consistently assessed inflammatory markers in relation to aSAH prognosis in the literature. The results show increase in baseline CRP was significantly associated with increased risk of DCI following aSAH, but not with poor functional outcome at follow-up. Increase in baseline WBC was not associated with DCI, however, were significantly associated with poor functional outcome.

Prediction of the occurrence of DCI and poor outcome is of critical value for aSAH management. It is known that the need for special laboratory test or collection of CSF sample limits the clinical utility of a biomarker, and this highlights the need for searching for a quick, sensitive and safe biomarker which is easy to obtain for evaluating prognosis in aSAH. In addition, there is increasing evidence on the role of inflammation in the pathogenesis of DCI and poor outcomes following aSAH [27]. Among the common readily available markers for systemic inflammation, many studies have investigated the independent role of early increase in CRP and WBC as risk factors for unfavorable outcomes after aSAH. However, the correlations remained inconclusive and there is no consensus on their value for assessing prognosis in aSAH. For example, early phase CRP after aSAH have predicted poor outcome and/or DCI in a few studies [14, 20, 23], while not in others [16, 24]. Conflict results were also observed with respect to studies on WBC [15, 26, 30]. On the other hand, there is an extensive body of literature identifying blood CRP and WBC as an independent and relatively strong predictive marker for outcomes after ischemic stroke [31,32,33], it is reasonable for expecting these associations in aSAH.

Suprisingly, this present meta-analysis found early increase in CRP is only associated with DCI but not later poor functional outcomes. There is a known association between DCI and worse clinical outcomes. However, sensitivity analysis showed that removing Sokół et al. [15] or Turner et al. [22] resulted in a significant association between poor outcome and CRP, suggesting the pooled estimate was overly impacted by these two studies. There was a relatively distinct definition of poor outcome (GOS 1) in Sokół et al. [15], and longer duration from hemorrhage onset to sampling in Turner et al. [22] than in other studies, which may help to explain the discrepancies.

Moderate to substantial heterogeneity was detected in the analyses of present review. A number of factors may be responsible for the heterogeneity. Firstly, the time point of outcome evaluation among the studies was not unified and not always clearly stated. Secondly, the aim of the present review was to analyze the associative role of early CRP/WBC. While most studies provided data of the first sampling on presentation, and although not markedly varied, there were still some variations in the timeframe of sample collection, and may contribute to the heterogeneity, since CRP is a sensitive marker and could have changed rapidly. Patient baseline characteristics, procedures performed, as well as factors adjusted for analyses differed across the studies, which may also contribute to the heterogeneity.

No meta-analyses have quantitively analyzed the associations between CRP/WBC and DCI/functional outcome following SAH, although several recent systematic reviews have provided qualitative assessment of a broad range of potential biomarkers for aSAH outcomes, including CRP and WBC [26, 34,35,36]. Jabbarli et al. [34] identified 21 genetic and non-genetic laboratory DCI biomarkers as ‘predictive biomarkers’ with level II evidence; peripheral CRP and WBC were evaluated in that review, of which level of evidence was considered insufficient, thus were allocated as ‘non-conclusive’ markers. Hong et al. [35] showed that serum CRP had increased specificity and sensitivity as a marker of inflammation after SAH and, as such, was a useful predictive indicator of vasospasm and less favorable functional outcomes after SAH. Authors of another study suggested utilizing CRP as prognostic indicator of SAH has a special value in sedated patients, because clinical condition of these patients is difficult to estimate [13].

The mechanisms underlying the associations observed remain unclear. Al-Mufti et al. [26] argued that WBCs produce reactive oxygen species (ROS) and disrupt the blood–brain barrier during extravasation into the cerebrum, which may contribute to DCI risk and poor outcomes in SAH. Those authors supported this by citing the finding of lymphocytic infiltrates on histopathologic staining of resected aneurysm tissue, which also have been considered as independent predictors of vasospasm [21, 30]. It is unlikely for us to infer whether or specifically how CRP/WBC may participate in the pathogenesis of DCI or the causal process in poor function after SAH. The observed associations may be explained by, as a previous study pointed out, that systemic inflammation as indicated by these parameters increases susceptibility to the development of DCI after SAH [16].

Authors of the review cited above also cautioned that CRP is an acute phase sensitive but non-specific inflammatory marker and it is not produced by brain tissue. SAH commonly has a secondary process such as pneumonia or other infection, which may lower the potential prognostic value of CRP [35]. In the present meta-analysis, differences in patients’ baseline characteristics and initial severity have been partially addressed by the multivariable models used in the included study adjusting for relevant factors (Table 2). Therefore, the pool estimation of CRP/WBC in related to outcome was not only a reflection of SAH severity. However, although several studies have, not all the included studies have excluded or adjusted for infection in their analyses. Also, assessment of infection is challenging and might have chance to be inaccurate [24]. Therefore, despite the fact we focused on the early period of aSAH and the parameters was mostly collected on admission in which infection was unlikely to present, the confounding by infection might have not been fully compensated. We are not able to make a further adjustment because of lacking raw data of individual studies.

Since CRP and WBC are relatively non-specific markers of inflammation. Evaluating other inflammatory markers such as Interleukin-6 (IL-6) and procalcitonin may provide additional insight. Unfortunately, few studies in the literature have looked into them; therefore, meta-analysis cannot be performed. Muroi et al. [24] investigated the relationships among IL-6, procalcitonin and outcomes after SAH, reporting a significant correlation between IL-6 and poor outcome (adjusted OR 4.03, P = 0.02) and DCI (adjusted OR 2.78, P = 0.03). Those authors concluded higher IL-6 levels might be a useful parameter to monitor. McMahon et al. [16] assessed the relation between IL-6 and DCI, while no significant association was found (adjusted odds ratio 1.1, P = 0.46). Another novel biomarker, red cell distribution width (RDW) which is strongly linked to inflammatory and thrombotic disease, was evaluated by three of the included studies [26, 29, 30]. Chugh et al. [30] reported a significant association between maximum daily RDW during first 10 days following hemorrhage onset and poor clinical outcome (adjusted OR 2.3, P = 0.014). Siegler et al. [29] also reported a significant association between elevated 14-day mean RDW and poor functional outcome at discharge (mRS > 4, adjusted OR 2.59, P = 0.040). However, elevated RDW in first 72 h was not associated with poor outcome in another study [26].

In the present review, we only identified associations and no predictive power was estimated due to insufficient reports. Among the studies included, Sokół et al. [15] reported the accuracy for prognosis of area under the curve (AUC) 0.74 (P < 0.05) in early CRP and poor outcome by calculating receiver operating characteristic (ROC) curve. Turner et al. [22] reported an AUC 0.779 in early CRP for predicting poor outcome at discharge among good grade patients. Al-Mufti et al. [26] reported combined utilizing Fisher grade and elevated admission WBC together had an AUC over 0.7, which was more accurate than Fisher grade alone. These reports together highlighted the potential predicting power of early CRP/WBC on aSAH outcomes, which is valuable to be confirmed in the future studies.

Strengths and limitations

This is the first meta-analysis to evaluate the role of peripheral CRP and WBC in outcomes of aSAH. The studies included were determined to be of good quality. This review has several limitations that should be considered when interpreting the results. Although fifteen studies were included, the number of studies included in the analyses for individual outcome were relatively small, which limited the statistical power. Initial severity of ictus, comorbidities, infectious status, surgery and external ventricular drain may largely influence the parameters studied, and were not fully compensated in the analyses. Another weakness of this meta-analysis is that only association was identified, no predictive power was estimated due to the lack of consistent reports. Inflammatory markers other than CRP/WBC, as well as different subtypes of WBCs were not analyzed due to insufficient data. Despite the described limitations, the present review give insight into the relationships of systemic inflammation with DCI and outcome after aSAH in general.

Conclusions

There appears to be some evidence indicating early increase in peripheral CRP is independently associated with the occurrence of DCI after SAH, and WBC associated with poor functional outcome at follow-up. Strong conclusion cannot be made due to the small study number, between-study heterogeneity and suspicion of uncontrolled factors. Future well-designed studies with prognostic accuracy evaluated, factors precisely adjusted and inflammatory biomarkers other than CRP/WBC assessed are highly warranted to determine whether monitoring circulating markers of inflammation benefits the risk identification after aSAH.