Prognostic Nutritional Index Predicts Clinical Outcome in Patients with Cerebral Venous Sinus Thrombosis


 Background: Lower prognostic nutritional index (PNI) is related to poor prognosis of cardiovascular disease. However, little is known about PNI and its relationship with prognosis in cerebral venous sinus thrombosis (CVST).Methods: From January 2013 to June 2019, we retrospectively identified consecutive CVST patients. We selected patients in acute / subacute phase as subjects. Poor prognosis was defined as modified Rankin Scale (mRS) of 3-6. Multivariate logistic regression analysis was used to confirm if lower PNI was associated with poor prognosis. Results: A total of 297 subjects with 12-month follow-up data were enrolled. Thirty-three (11.1%) had poor outcome. Multivariate logistic regression analysis suggested that PNI was an important predictive factor of poor outcome in acute/subacute CVST (odds ratio, 0.903; 95% CI, 0.833-0.978; P = 0.012). The optimal cut-off value for predicting a poor prognosis of PNI was 44.2. Kaplan-Meier analysis and log-rank test suggested that the lower the PNI value, the higher the mortality rate (P<0.001). In addition, the nomogram we set up showed that lower PNI was an index of poor prognosis. The c-indexes for the cute/subacute patients with CVST was 0.872.Conclusions: Lower PNI is correlated with a higher risk of adverse clinical outcome in patients with acute/subacute CVST.


Background
Cerebral venous sinus thrombosis (CVST) is a special type of cerebrovascular disease, which is characterized by cerebral venous re ux obstruction and intracranial hypertension caused by disturbance of cerebrospinal uid absorption. It is generally believed that the annual incidence of CVST is (2-5) / 1 million, accounting for 0.5%-1% of all strokes [1][2][3][4][5]. The inducing risk factors, clinical characteristics, and neuroimaging features of CVST are extremely diverse, which bring challenges to the prognosis [6].
Therefore, it is of important clinical signi cance to carry out early risk strati cation at the time of onset.
Malnutrition has been proven to be an independent prognostic index of incidence and mortality in patients with various cancers, myocardial infarction (MI) as well as acute ischemic stroke (AIS) [7][8][9].
The PNI is a novel and comprehensive nutritional-in ammatory score that re ects immune nutritional status on account of serum albumin concentration and lymphocyte count. Several studies have found that PNI is related to a poor prognosis and increased mortality in patients with malignant diseases, MI and AIS [10][11][12]. Nevertheless, this correlation between PNI and prognosis in CVST patients has not been studied yet. Acute/subacute CVST is characterized by sudden onset and rapid progression, which is of great signi cance to study [13]. Therefore, we aimed to assess the outcome and predictive role of PNI in patients with acute/subacute CVST.

Patient selection
From January of 2013 to June of 2019, we conducted a follow-up study at the Henan CVST Registry in the First A liated Hospital of Zhengzhou University (Henan, China). The diagnosis of CVST complied with international standards [4,5], based on clinical symptoms, patient history, and neuroimaging results (computed tomographic(CT), magnetic resonance venography (MRV) or digital subtraction angiography (DSA) ). We de ned exclusion criteria as follows: (1) patients without complete clinical data; (2) time from onset to admission over 30 days [13]; (3) patients less than 18 years old; (4) patients with severe hepatic or renal diseases; (5) patients lost to follow-up. The registry study was approved by the Ethics Committee of the First A liated Hospital of Zhengzhou University.

Data Collection
Demographic data and information such as age, gender, risk factors, clinical manifestations, laboratorial parameters and imageological examination were gathered and analyzed. Albumin concentration (ALB) and the absolute lymphocyte count (ALC) were assessed within 24h after admission.

Outcome Assessment
The outcome assessment was assessed by modi ed Rankin Scale (mRS) during routine follow-up which performed since discharge over the phone after three months. Adverse prognosis was de ned as mRS of 3-6. Telephone interviewers were unware of baseline data.

Statistical Analysis
Statistical analysis were performed using SPSS 22.0. Continuous data were represented as mean ± SD, which were analysed by independent Student's t-test or Mann-Whitney U test, whichever was appropriate.
Categorical variables were presented as numbers and percentage, which were compared by χ 2 test or Fisher exact test. Multivariate logistic regression analysis was used to analysis the association between PNI and clinical outcome. The receiver operating characteristic (ROC) analysis was estimated to evaluate the ability of the PNI in predicting clinical prognosis. We dichotomized the overall patients with the cut-off PNI value. The mortality rate in high and low PNI value groups was compared with the log-rank test and plotted with the use of Kaplan-Meier method. On this basis, a nomogram of independent predictive factors was established by using RMS software. Results were signi cant for Two-tailed P<0.05.

Study Subject Characteristics
Out of 324 patients who diagnosed with CVST within 30 days of onset to admission at this database were enrolled and 27 patients were excluded: 15 patients with younger than 18 years old, 8 patients without complete clinical data and 4 patients lost to follow-up. Therefore, 297 patients were included in the nal analysis. Comparison of baseline characteristics and treatment details of included and excluded patients, there was no observably otherness except the age (Table S1).

Characteristics of Patients with Good and Poor Prognosis
During the follow-up period, 33 patients (11.1%) had poor functional outcome, including 26 patients dead (8.8%). The baseline characteristics of the two groups were showed in Table 1. The PNI (49.3 ± 6.3 versus 42.3 ± 6.1, P < 0.001) of the poor outcome group was observably decreased. Patients with a poor outcome were older (34.2 ± 11.6 versus 44.4 ± 16.3; P < 0.001) compared with good outcome group. In addition, coma, focal neurological de cits, involvement of straight sinus, and intracerebral hemorrhage in neuroimaging manifestations were more common in patients with adverse prognosis. The lymphocyte count and albumin concentration in subjects with adverse outcome were obviously lower than those with good outcome (P < 0.001). Analysis Between PNI and Long-Term Prognosis As described in Table 2, multivariate logistic regression analysis suggested that PNI (odds ratio, 0.903; 95% CI, 0.833-0.978; P = 0.012) remained signi cantly correlated with the functional outcome after the adjustment for potential confounders including age, gender, presence of coma, focal neurological de cits, intracerebral hemorrhage, and straight sinus. According to the ROC curve of laboratory parameters to predict poor results showed in Fig. 1, when 44.2 was set as the cut-off point, PNI showed the best predictive value with the receiver operating characteristic analysis and area under the curve (AUC) of 0.75 (sensitivity: 75.8% and speci city: 69.7%), which showed the highest AUC value than those of ALC (AUC 0.71; 95% CI, 0.63-0.79) and ALB (AUC 0.69; 95% CI, 0.59-0.79). It indicated that PNI provides a stronger predictive power of poor outcome than its constituent parts. The mRS distribution was shown in Fig. 2.
Overall, patients with lower PNI had worse scores on the mRS than higher PNI patients. Kaplan-Meier analysis and log-rank test suggested that the lower the PNI value, the higher the mortality rate (P < 0.001) (Fig. 3).

Nomograms to Predict the Prognosis
In addition, age, coma, intracranial hemorrhage, straight sinus involvement and PNI were included in the nomogram of the study (Fig. 4) by stepwise logistic regression analysis. The new model showed that older age, coma, intracranial hemorrhage, straight sinus involved and lower PNI were indicators of poor prognosis. The c-indexes of the model was 0.872. These ndings were similar to those of the previous multivariate logistic model.

Discussion
In the International Study on Cerebral Vein and Dural Sinus Thrombosis (ISCVT), male gender, age > 37 years old, mental status disturbance, coma, thrombosis of the deep venous system, intracranial hemorrhage, malignancy, and infection of the central nervous system were independently associated with adverse outcome at last follow-up [15]. Our study showed that PNI was a signi cant and independent predictor of poor prognosis in non-chronic CVST patients.
Current research supports the link between in ammation and the incidence of CVST [16,17]. A number of studies appeared to reinforce the biological plausibility behind in ammation and the prognosis of cerebral venous thrombosis [18,19]. The lymphocyte count is an indicator that mediates cellular immunity. It is worth noting that some studies supported the use of lymphocyte count as a prognostic indicator. Lymphocytes are involved in cellular immunity of various cancers and are related to cancer prognosis [20]. In addition, previous study has shown that lymphocyte count is a predictive factor of adverse outcome in patients with AIS [21]. In animal models, immunosuppression induced by stroke could lead to lymphopenia and changes in the ratio of helper T cell [22,23]. Similarly in human studies, it had also been observed that peripheral blood lymphocytes decreased after stroke, particular in the acute phase [24,25]. Therefore, lower lymphocyte count may be a predictor of adverse outcome. The relationship between lymphocyte count and the outcome of CVST has been con rmed in previous studies, such as platelet to lymphocyte ratio (PLR), lymphocyte to monocyte ratio (LMR) and neutrophil to lymphocyte ratio (NLR) [26][27][28].
Hypoalbuminemia is a comprehensive result of in ammation and insu cient intake of protein and calories in patients with chronic diseases. In different clinical settings, hypoalbuminemia has been shown to be a sign of poor prognosis [29][30][31]. Considering that albumin is a negative acute phase protein, its synthesis rate is affected by nutrition and in ammation [32]. More and more evidence shows that as the severity of in ammation increases, serum albumin levels gradually decrease [33,34]. Albumin combined with nitric oxide (NO) free radicals has anticoagulant and antithrombotic effects. Due to the increase in the concentration of free lysophospholipids, hypoalbuminemia may affect blood viscosity and endothelial cell function [35]. However, serum albumin is not a good indicator of nutrition observation because it is greatly affected by uid transfer [29]. Many different factors affect serum albumin levels and have shown a lack of sensitive and speci c indicators of nutritional status [36].
As for PNI, it combines the lymphocyte count and albumin concentration, re ecting the nutrition, in ammation and immunity status. Therefore, compared with the above single factors, PNI is more stable and representative. The PNI was originally reported to be used to evaluate the immune and nutritional status of patients undergoing gastrointestinal tract surgery [14,37,38]. Subsequently, PNI has been widely used in prognostic evaluation of a variety of cancers and transplant operations, as well as for patients with various diseases such as myocardial infarction, acute type A aortic dissections and AIS patients receiving intravenous thrombolysis (IVT) [10,11,39,40].
This study has further shown that lower PNI increased the risk of poor outcome in patients attributable to acute/subacute patients with CVST. Additionally, the ndings from nomograms by stepwise logistic regression analyses suggested that age, coma, intracerebral hemorrhage, straight sinus involvement and PNI were also predictors of adverse outcome in acute/subacute patients with CVST, which further supported the results from multivariate logistic regression analyses. Therefore, the PNI created by combining serum albumin concentration and lymphocyte counts have the ability to assess the nutritional, in ammatory and immune status of patients with acute/subacute CVST. Considering that patients with lower PNI scores in this study have a signi cantly higher incidence of poor prognosis, it can be considered that appropriate evaluation and implementation of measures to improve nutritional status will help improve the outcome of patients with acute/subacute CVST. Further prospective studies are required to verify this hypothesis.
PNI is easy to obtain because it is calculated using objective laboratory test data. This makes it easy for our achievements to be translated into daily practice. However, several potential limitations should be acknowledged in the present study. First, it is a single-centre study and the sample size included in our study is relatively small. This association in the present study also needs to be con rmed and veri ed in larger multicenter prospective cohort studies. Second, compared with arterial stroke, it is di cult to determine the exact time of onset in patients with CVST.
In summary, our study suggested that lower PNI was a potential dangerous factor in unfavorable functional outcome of patients with acute/subacute CVST.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. TableS1.docx