D-dimer may predict poor outcomes in patients with aneurysmal subarachnoid hemorrhage: a retrospective study.

Serum biomarkers may play a reliable role in predicting the outcomes of patients with aneurysmal subarachnoid hemorrhage. This study retrospectively analyzed the relationship between serum biomarkers on admission and outcomes in patients with aneurysmal subarachnoid hemorrhage. We recruited 146 patients with aneurysmal subarachnoid hemorrhage who were treated in Renmin Hospital of Wuhan University of China between 1 May 2014 and 30 March 2016. There were 57 males and 89 females included and average age of included patients was 57.03 years old. Serum samples were taken immediately on admission (within 48 hours after initial hemorrhage) and the levels of serum biomarkers were detected. Baseline information, complications, and outcomes at 6 months were recorded. Univariate and multivariate logistic regression analyses were used to explore the relationship between biomarkers and clinical outcomes. Receiver operating characteristic curves were obtained to investigate the possibility of the biomarkers predicting prognosis. Of the 146 patients, 102 patients achieved good outcomes and 44 patients had poor outcomes. Univariate and multivariate analyses showed that high World Federation of Neurosurgical Societies grade, high serum D-dimer levels, and high neurological complications were significantly associated with poor outcomes. Receiver operating characteristic curves verified that D-dimer levels were associated with poor outcomes. D-dimer levels strongly correlated with neurological complications. In conclusion, we suggest that D-dimer levels are a good independent prognostic factor for poor outcomes in patients with aneurysmal subarachnoid hemorrhage.

Introduction

Aneurysmal subarachnoid hemorrhage (aSAH) is a devastating disease with an annual incidence of 2–25 per 100,000 persons (de Rooij et al., 2007; Liu et al., 2011; Takashima et al., 2017; Wang et al., 2017). Although the management of aSAH in the past two decades has rapidly developed, 24.8% of patients still die in the first year after initial hemorrhage and more than 40% of survivors have various degrees of disability (de Azúa López Zaida et al., 2015; Mukhtar et al., 2016). Several clinical studies had demonstrated that age, Hunt-hess grade, World Federation of Neurosurgical Societies (WFNS) scale grade, and Fisher grade on admission are associated with unfavorable outcomes in patients with aSAH (Rosengart et al., 2007; Jaja et al., 2015; Zhao et al., 2017). In spite of these clinical factors, it is still necessary for clinicians to find more functional outcome predictors because predicting outcomes solely based on clinical factors are sometimes inaccurate (Le Roux et al., 1996; Turner et al., 2015; Greenberg et al., 2017).

Biomarkers are easy to obtain and measure, and because of their potential role in the pathological process after aSAH, serum biomarkers might be more reliable in predicting the outcomes of patients with aSAH (Hong et al., 2014; Ding et al., 2016; Paschoal et al., 2016; Suzuki and Kawakita, 2016; Frontera et al., 2017). Clinical studies regarding the relationship between these biomarkers on admission and outcomes are limited and more studies are needed to verify the roles of serum biomarkers in predicting the prognosis of patients after aSAH. Numerous factors might lead to elevated levels of D-dimer after aSAH, such as enhanced fibrinolysis activity, disruption of vessels, tissue factor release, and dissolved clots (Larsen et al., 2012; Staykov and Schwab, 2013; Boluijt et al., 2015). Clinical studies have revealed that D-dimer levels are elevated in patients with aSAH (Parra, 2006). Few studies have investigated the relationship between D-dimer levels and poor outcomes, and their prognostic role in aSAH. Thus, this study aimed to explore the prognostic role of D-dimer levels in patients with aSAH in the hope the results might give some guidance to the management in aSAH.

Subjects and Methods

Subjects

A retrospective study was conducted in aSAH patients, who were treated in Renmin Hospital of Wuhan University of China between 1 May 2014 and 30 March 2016. Patients included were diagnosed with aSAH by head computed tomography (CT) and head CT angiography (CTA) (Ni et al., 2016) or digital subtraction angiography (DSA) (Bechan et al., 2015). The aSAH patients also received hemostatic agents, mannitol or furosemide, nimodipine, prophylactic anti-seizure drugs, antihypertensive drugs, and other symptomatic treatment if necessary immediately on admission. After patients were diagnosed with aSAH, they received either microsurgical clipping or coiling. After surgery, patients also received standard treatments, as well as head CT and CTA. An immediate CT scan was performed to confirm any pathological changes if patients had neurological deterioration.

Inclusion criteria: Patients presenting with all of the following criteria were considered for study inclusion: (1) aSAH by head CT and head CTA or DSA (Grasso et al., 2017); (2) time from initial hemorrhage to admission < 48 hours. Exclusion criteria: Patients with one or more of the following conditions were excluded from this study: (1) perimesencephalic nonaneurysmal SAH or SAH caused by head trauma, vascular anomaly, vascular malformation, or moyamoya disease; (2) bilateral mydriasis, dilated pupils, or other irreversible brainstem injury; (3) severe heart/renal/respiratory/liver dysfunction; (4) taking anticoagulant drugs or corticosteroids or infection diseases in the 1 month before hospitalization. The study followed the principles of the Declaration of Helsinki and was approved by the Medical Ethics Committee of the Renmin Hospital of Wuhan University of China.

Data collection

We recorded the details (name, sex, and age) of patients and asked patients or their relatives about the history of the disease process. Details of accompanying diseases were acquired, as well as history of smoking and drinking. The Glasgow Coma scale (GCS) (Green et al., 2017) and WFNS Scale (Sano et al., 2015) were used to assess neurological status. Patients were categorized by Fisher grade (Smith et al., 2005) according to the head CT presentations on admission. Aneurysm location and the size of the neck of aneurysm were also recorded. Patients received either surgical clipping or coiling as soon as possible. We recorded postoperative complications, including systematic complications and neurological complications. Neurological complications included delayed cerebral ischemia, hydrocephalus, and seizure. Delayed cerebral ischemia was defined as an unexplained neurological deterioration or a new infarct on head CT after surgery, but not surgery-related cerebral ischemia (Suarez, 2015; Foreman, 2016). Systematic complications were defined as pneumonia, urinary tract infection, or sepsis (Zheng and Wong, 2017). The Glasgow outcome scale (GOS) was used to assess the neurological function at 6 months after the initial hemorrhage (Turek et al., 2016). A favorable outcome was defined as 3–5 on the GOS and 1–2 on the GOS was defined as poor outcome (Giraldo et al., 2012; Elhadi et al., 2015).

Serum biomarker measurement

Serum samples were taken immediately on admission (within 48 hours after initial hemorrhage) and sent to clinical laboratory of Renmin Hospital of Wuhan University for further analysis. Standard quantitative assay techniques were used to detect the levels of serum biomarkers and all operations were consistent with manufacturers’ instructions (AU5800, Beckman, CA, USA). Normal values of the biomarkers were as follows: leukocytes: 4–10 × 109/L, blood platelets: 100–300 × 109/L, K+: 3.5–5.5 mM, Na+: 135–145 mM, D-dimers: 0–0.55 mg/L, C-reactive protein: 0–10 mg/L, lactic dehydrogenase: 135.0–215.0 U/L, and low density lipoprotein-cholesterol: < 3.1 mM (Rahmanian et al., 2015).

Statistical analysis

Descriptive analysis (mean, standard deviation, and median) was used for continuous variables and percentages for categorical variables. The analysis of the association between variables (baseline information and biomarkers on admission) and clinical outcome was evaluated using univariate logistic regression analysis and multivariate logistic regression was used for further analysis. If variables were P < 0.05, receiver operating characteristic (ROC) curves were used to access the prediction ability of the variants. We conducted additional analyses of candidate biomarkers and complications using logistic regression analysis and also explored the correlations between candidate biomarkers and baseline information. Correlations between variables were analyzed using Spearman's rank correlation test. Statistical analyses were conducted using SPSS 21.0 software (IBM, Armork, NY, USA) and Medcalc (Medcalc, Ostend, Belgium). P < 0.05 was considered statistically significant.

Results

Characteristics of aSAH patients

One hundred and forty-six patients with aSAH were included in this retrospective study, including 57 males and 89 females. One hundred and two of these patients achieved good outcomes at 6 months after initial hemorrhage, whereas 44 patients had poor outcomes. The mean ages of patients with good outcomes and poor outcomes were 56.68 ± 8.16 and 57.86 ± 9.92 years old, respectively. There were 50 patients with hypertension, 16 with diabetes, 27 patients who smoked. GCS scores were 3–8 in 31 patients, 9–11 in 17 patients, and 12–15 in 100 patients on admission. There were 101 patients with a WFNS grade of I–III and 45 patients with a WFNS grade of IV–V. Hunt and Hess scale grade was used to access the status of patients (Ghosh et al., 2012) with grades I–II in 80 patients and grades III–IV in 66 patients. The mean time to treatment commencement (initial hemorrhage to admission) was 12.78 hours. One hundred and eleven aSAH patients received neurosurgical clipping and 35 patients received endovascular coiling. Patients’ baseline details are shown in .

Table 1

Baseline information of 146 patients with aneurysmal subarachnoid hemorrhage

Postoperative complications

Both neurological and systematic complications were recorded in this study. The results showed that 18 patients suffered from delayed cerebral ischemia, 19 patients experienced seizures, and 15 patients developed chronic hydrocephalus. Thirty-four patients had systematic complications, including fever, infections, and abnormal renal function. Details are shown in .

Table 2

Postoperative complications of survival patients with aneurysmal subarachnoid hemorrhage

Analyses of variables predicting poor outcomes

Results of the univariate and multivariate analyses are shown in Tables and , respectively. WFNS grades IV–V, Fisher grades III–IV, D-dimer levels, C-reactive protein, systematic complications, and neurological complications were associated with poor outcomes (P < 0.05). We selected three main biomarkers (P < 0.05), which were associated with poor outcomes in univariate analyses to perform multivariate analysis. These results showed that WFNS grade and D-dimer levels were significantly associated with poor outcomes ().

Table 3

Univariate analysis for variables of a poor outcome

Table 4

Multivariate analysis of variables indicating a poor outcome

Prediction ability of biomarkers

We performed ROC to assess each biomarker's prediction ability. The areas under the curve were 0.69, 0.86, 0.61, and 0.71 for C-reactive protein, D-dimers, leukocytes, and WFNS grade IV–V, respectively. The sensitivity of C-reactive protein, D-dimer, leukocytes, and WFNS IV–V in predicting poor outcomes was 70.45%, 82.09%, 50.00%, and 59.09%, respectively. The specificity of these four variables was 65.69%, 78.43%, 76.47% and 82.35%, respectively. Details are presented in .

Figure 1

Diagnostic value of biomarkers for poor outcomes.

CRP: C-reactive protein; WBC: white blood cell; WFNS: World Federation of Neurosurgical Societies.

Additional analysis of D-dimers

D-dimer levels were positively correlated with high WFNS grade, high Hunt-hess grade, and high Fisher grade, and were also strongly correlated with neurological complications (). Univariate and multivariate regression analyses were used to evaluate the contribution of D-dimer levels to complications. The results showed that D-dimer levels were significantly associated with neurological complications (OR = 2.56, 95%CI (1.47–3.32), P < 0.001), but not systematic complications (OR = 1.10, 95%CI (0.85–1.43), P = 0.46). Details are shown in .

Table 5

Results of Spearmen's correlation between variables and D-dimer

Table 6

Multivariate analysis for contribution of D-dimer to complications

Discussion

This was a retrospective study in which the patients were diagnosed with aSAH and underwent medical treatment within 48 hours. Preliminary findings showed that WFNS grades IV–V, Hunt and Hess scale grades III–V, Fisher grades III–IV, D-dimers, C-reactive protein, leukocytes, systematic complications, and neurological complications were associated with poor outcomes (Orakdogen et al., 2016; Aggarwal et al., 2017; Dinc et al., 2017). Several studies have demonstrated that higher WFNS, Fisher, and Hunt and Hess scale grades are associated with poorer outcomes in patients with aSAH (Sano et al., 2016; Zhao et al., 2016; Kilic et al., 2017; Ljubisavljevic et al., 2017). Moreover, a predictable model built by Zhao et al. (2017), which combined baseline information and postoperative pneumonia, had excellent discrimination. WFNS grade V and Fisher grades III–V were found to be independent factors for a poor outcome (OR = 3.728, 6.234, respectively). Moreover, they had excellent prognostic value in discriminating poor and good outcomes with an area under curve of 0.87 (van Donkelaar et al., 2017). Biomarkers, such as C-reactive protein, interleukin-6 (Ljubisavljevic et al., 2017), serum tumor necrosis factor-α (Sarrafzadeh et al., 2010; Sun et al., 2017), and erythrocyte distribution (Chugh et al., 2015), were all found to be associated with poor outcomes in aSAH patients. Inflammatory factors on admission, including high-sensitivity-C-reactive protein and leukocyte levels, were shown to have independent associations with GOS scores at 3 months (Srinivasan et al., 2016). The level of C-reactive protein on admission failed to show a significant association with poor outcomes, as well as neurological and systematic complications in this retrospective study (Romero et al., 2012; Badjatia et al., 2015). In this study, leukocyte levels were obviously higher in patients with poor outcomes, but it was not an independent risk factor for poor outcome. Leukocyte levels were also not a good prognostic factor for poor outcomes. This result was the same as a previous study (Rothoerl et al., 2006), where leukocytes had no significant relationship with clinical outcomes.

D-dimer is a fibrin degradation product and has been investigated in many diseases, such as deep vein thrombosis, acute aortic dissection, intracerebral hemorrhage, and also aSAH (Fujii et al., 1995; Suzuki et al., 1997; Ramchand et al., 2016; Fukuda et al., 2017). Patients with poor outcomes had higher D-dimer levels, which predicted poor outcomes independently. D-dimer levels were significantly higher in patients with SAH and were first found to be correlated with poor outcomes in 1995 by Fujii et al. Several studies have demonstrated that D-dimers are an independent predictor for delayed ischemic neurological deficit and delayed cerebral ischemia in patients with aSAH. In our study, D-dimer levels showed significant correlation with neurological complications and were also significantly associated with neurological complications, which differs from previous results (Fukuda et al., 2017). It is believed that pathophysiological processes after SAH are complicated and that certain factors might play different roles at different times (de Lima Oliveira et al., 2015; van Lieshout et al., 2017). The dynamic changes of D-dimer levels, which were not investigated in our study, are very important. Ilveskero et al. (2005) found that D-dimer levels on admission and after surgery were elevated and correlated with long-term outcomes. In this retrospective study, we confirmed that D-dimer levels on admission had significant association with poor outcomes. Furthermore, the area under the curve of D-dimers in ROC was 0.86, which indicated that it had good prognostic value in predicting poor outcomes. Similar findings were reported in a previous study and the results showed that D-dimer levels after admission could predict poor outcomes independently (Juvela and Siironen, 2006).

Elevation of D-dimer levels always indicates enhanced fibrinolysis activity in the human body and is a biomarker of a hypercoagulative state in patients with SAH (Filizzolo et al., 1978; Burchiel et al., 1984; Ilveskero et al., 2005). The possible mechanism of elevated plasma D-dimer levels might be as follows: The sudden disruption of vessels causes injury to the endothelium and subsequently releases tissue factor. Tissue factor release and dissolved subarachnoid clots may play important roles in fibronolysis, which led to D-dimer elevation (Fukuda et al., 2017). However, in previous studies (Mocco et al., 2006; Boluijt et al., 2015), no association was found between D-dimers on admission and delayed cerebral ischemia, which differs from the results in our study. Our results confirmed that elevated D-dimer levels are significantly associated with neurological complications (including delayed cerebral ischemia and seizures). Moreover, elevated D-dimer levels could be used as an independent prognostic factor for neurological complications.

In this study, we only investigated preoperative levels of D-dimer and this indicated that D-dimers have a significant association with poor outcomes with good sensitivity and specificity in predicting poor outcomes in patients with aSAH. Blood samples were collected on admission, because many factors, such as surgery and deep venous thrombosis, might have effects on D-dimer levels after surgery (Delgado et al., 2006). Furthermore, preoperative D-dimer levels might provide more reliable evidence for fibrinolytic activation caused by the rupture of intracranial aneurysm, which might contribute to early brain damage and in-hospital complications. Elevated D-dimer levels on admission provide information about whether patients might achieve good outcomes or poor outcomes, as early as possible. In this study, we confirmed that D-dimer levels could be used as a prognostic factor for poor outcomes in aSAH patients. However, the real role of D-dimer levels as a prognostic factor for patients with aSAH is still unclear and needs further clinical studies, with strict inclusion and exclusion criteria, to explore its dynamic change before and after surgery.

Several limitations exist in this study. We did not explore the dynamic change of D-dimer levels in patients with aSAH and the number of patients was relatively small. D-dimer levels may differ between patients who received neurological clipping and underwent endovascular coiling, and further studies should focus on each of these groups.

In summary, plasma D-dimer levels are significantly elevated in aSAH patients with poor outcomes at 6 months after the initial hemorrhage. D-dimer levels had a significant association with poor outcomes and neurological complications. D-dimers levels are good and independent prognostic factors for poor outcomes in aSAH patients.