Clinical Predictive Factors of Lower Extremity Deep Vein Thrombosis in Relative High-Risk Patients after Neurosurgery: A Retrospective Study

Introduction It is acknowledged that patients undergoing neurosurgery with neurological illness are at higher risk of lower extremity deep vein thrombosis (DVT). As an underlying life-threatening complication, the incidence and risk factors for high-risk patients with lower extremity deep vein thrombosis are still controversial in relative high-risk patients after neurosurgery. Materials and Methods A total of 204 patients who underwent neurosurgery and were considered as a high-risk group of DVT according to times of stay in bed more than 3 days were enrolled in this study. We evaluated the lower extremity DVT by using Color Doppler Ultrasound System (CDUS). Clinical parameters of patients at the time of admission and postoperation were recorded and prepared for further analysis. Early predictive factors for postoperative lower extremity DVT were established. Diagnostic performance of predictive factors was evaluated by using receiver operating characteristic (ROC) curve analysis. Results The overall incidence rate of DVT in 204 enrolled patients was 30.9%. Multivariate logistic regression indicated that hypertension (OR 3.159, 95% CI 1.465-6.816; P = 0.003), higher postoperative D-dimer (OR 1.225, 95% CI 1.016-1.477; P = 0.034), female (OR 0.174, 95% CI 0.054-0.568; P = 0.004), and lower GCS score (OR 0.809, 95% CI 0.679-0.965; P = 0.013) were independently associated with incidence of DVT in patients after neurosurgery. The logistic regression function (LR model) of these four independent risk factors had a better performance on diagnostic value of DVT in patients after neurosurgery. Conclusion The combined factor was constructed by hypertension, postoperative D-dimer, gender, and GCS score, and it might be a more handy and reliable marker to stratify patients at risk of DVT after neurosurgery.


Introduction
Deep vein thrombosis (DVT) is defined as formation of a thrombus within a deep vein and often caused by hypercoagulation and vascular endothelial injury [1]. Each year, over 550,000 hospitalized patients were affected by DVT in the United States, and DVT might cause a serious financial burden on patients [2,3]. DVT is considered a common vascular disease in patients and accounts for 30% cases of pulmonary embolism (PE) [4]. As a life-threatening complication, mortality rate in patients with pulmonary thrombosis is about 14% [5]. It is important to recognize patients at risk for DVT early as they may require antithrombotic and thrombolytic therapy.
For surgical patients, those who undergo neurosurgery are considered to be much more susceptible to DVT [6]. Statistics suggest that incidence of venous thrombosis complications in neurosurgery patients varies from 1.7 to 34%; researches further indicated that brain malignant tumor and limb plegia of prolonged duration could increase the risk of venous thrombosis [7][8][9]. Patients with DVT usually present with a mild form. However, a higher overall mortality in about 6% of patients with DVT was recorded. While about 30% of patients with DVT develop into pulmonary embolism, overall mortality even exceeds 10% [4,10]. As an effective measure, early initiation of heparin treatment is recommended for prevention of DVT [11]; however, it is still controversial, especially in patients with a higher risk of bleeding. Although several previous studies have identified that old age, restricted movement, overweight, cancer, pregnancy, and surgery are risk factors for DVT, the simple and accurate biomarkers for DVT in patients are still needed. Early screening of the DVT in neurosurgery patients is very necessary for lower morbidity and mortality. Evaluation of risk factors of DVT for neurosurgery patients is necessary for postoperative more aggressive mechanical and chemical prophylaxis.

Patients and Study
Design. This is a retrospective observational study and approved by the Institutional Review Board of the First Affiliated Hospital of Nanjing Medical University; patients who underwent neurosurgery admitted to the First Affiliated Hospital of Nanjing Medical University (Nanjing City, Jiangsu Province, China) from January 1 to December 31, 2018, were included in the study cohort. Informed consent from individual participants was obtained. As shown in Figure 1, of 742 patients, 538 patients were excluded from the study cohort for one or more of the following reasons: (1) age < 18 years old; (2) incomplete medical records; (3) history of anticoagulant therapy; (4) vascular disease surgery history; (5) history of atrial fibrillation; (6) history of lower extremity venous thrombosis; (7) history of malignant tumor; and (8) postoperative bed rest more than 3 days. Utilization of mechanical and physical therapy was applied to avoid DVT except for anticoagulant therapy in patients before or after surgery. Relative high-risk patients were defined as people who had bed rest more than 3 days after neurosurgery.
All enrolled patients were grouped into a DVT group or non-DVT group according to whether there is a DVT by duplex ultrasonography examination. All clinical characteristics of enrolled patients such as gender, age, BMI (body mass index), hypertension, pathological type, bedridden time, past medical history, and preoperative and postoperative D-dimer serum levels were retrieved from medical records. An electronic health medical record system and paper-based medical record library which allows access to more detailed information were noted in this study.

Diagnosis of DVT.
Autar DVT risk assessment scale was an accurate and reliable indicator for DVT，and it was applied for the risk assessment of thrombosis in both groups on the day before surgery. Both lower extremities were confirmed with DUS (Vinno, Suzhou, China) by two different experienced ultrasound doctors preoperation and on postoperative days 1, 3, 6, 9, and 14.
2.3. Statistical Analysis. All data were performed in SPSS17.0 software (SPSS Inc., IBM Corporation, USA) and Stata/SE version 10.0 for Windows. Results were presented as mean ± standard deviation for continuous variables and frequencies (percentages) for categorical variables. The comparison of continuous variables was performed by Student's t-test between the two groups, and chi-squared test or Fisher's exact test was used to compare categorical variables. Univariate logistic regression models were used to examine the association between potential risk factors and DVT, and the multivariate logistic regression analysis was performed to determine the independent risk factors for DVT in  Disease Markers neurosurgery patients. A two-tailed P value less than 0.05 was used to indicate statistical significance. Receiver operating characteristic (ROC) curve analysis was performed to examine the predictive role of factors for the diagnosis of DVT and to determine the cutoff values.

Univariate Logistic Regressions.
To assess the independent predictors of DVT in patients after neurosurgery, univariate logistic regression analysis was performed and the results (

Discussion
Currently, it has been reported that postoperative patients are at significantly increased risk of DVT especially in neuro-surgical patients. Prior studies had suggested that incidence of DVT in neurosurgical patients varies from 6% to 43% [12]. As a result of lacking specific symptoms, DVT is often ignored and misdiagnosed in the clinical work. DVT is an important cause of prolonged inpatient stays and unexpected death in hospitalized patients; however it is considered a preventable complication. DVT of the lower extremity in neurosurgical patients can be caused by various factors both in genetic and in acquired factors [13,14]. Our study was designed to investigate the incidence of DVT and identify potential predictors for DVT in patients after neurosurgery.
In neurosurgical patients, there were still no general guidelines and uniform standards for preventing DVT, though nursing mechanical intervention and pharmacologic anticoagulant treatment have been confirmed to effectively reduce DVT. However, pharmacologic prophylaxis such as heparin is not widely accepted in neurosurgical operation because of possible intracranial hemorrhage. How to select early anticoagulant treatment and nursing mechanical intervention on relative high-risk patients after neurosurgery is still unclear. According to the recent American College of Chest Physicians (ACCP) 2012 guidelines, mechanical prophylaxis is preferred  Disease Markers over pharmacological prophylaxis in routine craniotomy patients. The most common mechanical methods for DVT prophylaxis are compression stockings (CS) and intermittent pneumatic compression (IPC) of the lower extremities [15]. This study was conducted to screen strategy for the detection of DVT in neurosurgical patients, and a more targeted preventive measure should be taken for relative high-risk groups.
The findings of this study showed that the incidence of DVT was 30.9% in patients who kept in beds more than 3 days after neurosurgery during hospitalization. Univariate and multivariate analysis results indicated that female correlated with DVT risk. Previous studies revealed that men have a slightly higher overall incidence rate than women, whereas during the reproductive years, women have a slightly higher rate [16]. An appearance of a hypofibrinolytic state was observed in pregnancy period. At that time, coagulation factors were increased including fibrinogen, factor VIII, and factor von Willebrand and inhibitors such as protein S were decreased, while taking hormonal contraceptives and hormone replacement therapy can also increase the risk of DVT in women after neurosurgery. Further information of enrolled women should be gathered for further analysis.
Despite the use of serum D-dimer, it seemed to be a reliable marker for the detection of DVT. But a variety of conditions might increase the D-dimer level including hemorrhage, infection, and trauma; what is more, there is no set value of D-dimer that implies the occurrence of DVT. In our study, despite the fact that both preoperative D-dimer and postoperative D-dimer were statistically higher in DVT-positive patients, only postoperative D-dimer was an independent risk factor for DVT in patients after neurosurgery. Our results indicated that a high value of preoperative D-dimer does not indicate the existence of DVT. Detection value of postoperative D-dimer is important.
The results presented here also found that a GCS score less than 5 was an independent risk factor for DVT. GCS includes eye opening, verbal, and motor responses, which was used ubiquitously in acute care databases and serves as an assessment tool for acute neurologic injury [17][18][19]. The present study suggested that a lower GCS score should be paid more attention in patients after neurosurgery. Multivariable models suggested that hypertension was also an independent risk factor of DVT in neurosurgical patients. Hypertension has been verified as a risk factor of DVT previously [20]. Several prospective studies have suggested that hypertension patients have a 2-fold increased likelihood of developing DVT [21]. We came to the same conclusion in neurosurgery patients.
Autar score was widely used as a clinical scoring system for DVT, However, this scoring system was not only for neurosurgery patients, and a large number of parameters were required for calculation, which makes it clinically cumbersome. We kept the four variables (hypertension, postoperative D-dimer, gender, and GCS) in a new model for prediction of DVT in neurosurgery patients. The final parameter calculation was based on regression coefficients. Based on the ROC curve analysis, the combined factors achieved higher AUC, compared with Autar score performed. We considered that it was easy to use for prediction of DVT in neurosurgery patients.
There were also some potential limitations that should be acknowledged in this study, such as relatively small sample size, single-center study, and retrospective nature. These limitations might lead to several biases and false conclusions. In addition, a well-designed, prospective study with more data will be needed to validate the results of this study.

Conclusion
In conclusion, we found that hypertension, postoperative Ddimer, gender, and GCS score were independently associated with incidence of DVT in patients after neurosurgery. Compared to Autar score, the combination of these four independent risk factors had a better performance on the diagnostic value of DVT in patients after neurosurgery, and it could be a more handy and reliable marker to improve clinical care and management strategies in patients after neurosurgery. Meanwhile, additional studies and statistics are required to further validate our results.

Data Availability
All data in our study are available from the corresponding authors upon reasonable request.

Ethical Approval
This study was approved by the institutional review board at the First Affiliated Hospital of Nanjing Medical University. Informed consent from individual participants was obtained.

Conflicts of Interest
We declare that we have no conflict of interest.