Assessment of anesthesia and surgical risks, which are also closely related to prognosis, are important for patients undergoing surgery. Currently, GCS score is commonly used to predict the prognosis in neurosurgery [4]. A study with 27625 cases of craniofacial trauma indicated that when the three elements in the GCS scoring system is analyzed separately, they could serve as independent predictors for the prognosis, when the three components are evaluated together, they could predict the mortality of these patients [5]. For patients with severe craniocerebral trauma, factors such as preoperative GCS score, age and body temperature are found to be significant predictors for prognosis [6]. Analysis on 305 cases of severe craniocerebral trauma patients with GCS scores of ≤ 8 showed that when the patient's age is less than 50 years, GCS is high and ICP < 20 mmHg, the mortality is low and the prognosis is relatively good [7]. In this study, although the GCS score is found to be correlated to prognosis, it is not an independent risk factor affecting prognosis.
Currently, ASA and APACHEE II scores are widely used criteria for assessing patients' general conditions. ASA class is mainly used to assess the anesthesia risks in perioperative periods. In addition, ASA class is also shown be able to predict prognosis after surgery [8, 9]; the correlation analysis in 6301 patients between ASA class and prognosis showed that ASA class predicts prognosis [10]. For example, for 22600 patients undergoing total hip arthroplasty and 18434 patients with knee arthroplasty in New Zealand between 2005 and 2008, ASA class can be used to predict postoperative mortality and functional status, as well as the early failure rate of surgery [11]. However, for neurosurgery, whether ASA class predicts prognosis is still debating [12]. This may be due to the fact that ASA classification is relatively subjective, and is, therefore, dependent on the training and experience of physicians [13]. In addition, regional and professional variations in ASA classification are also evident [14–16]. Therefore, more objective, reliable and consistent assessments with clinical operability are very desirable [17].
APACHE II scoring system is one of the widely used assessments for severe illness in ICU [18]. Raj et al. found that it can predict 6 month-mortality well but not prognosis for ICU patients with severe craniocerebral trauma [19]. In another study, APACHE II score at discharge was found to be able to predict the readmission to neurosurgery ICU, when the score is > 8.5, the patients may be suggested to extend ICU stay to reduce the readmission risk [15]. Compared with GCS score, APACHE II score was found to have better sensitivity and specificity in predicting the mortality of patients with brain traumatic injury [16] and if APACHE II score is ≥ 15, it predicts poor prognosis and low mortality [20]. These studies confirm that APACHE II score is reliable for predicting mortality. However, the predictability was believed to be due to the use of the worst value of several physiological variables in the first 24 hours in ICU. Any score that uses data collected over 24 hours is affected by the quality of care provided [21, 22]. Our results also indicate that APACHE II score is significantly related to the outcome, but the OR is low as pointed out previously [23].
Anesthesia class used in the study is based on combined assessment of preoperative conditions including ASA class and other patient's information such as age and the score was shown to have significantly value in predicting perioperative mortality [24]. Binary logistic regression analysis showed that anesthesia class has an OR of 3.933, which is higher than these of ASA score and APACHE II class.
Although the study generates useful information regarding prognosis prediction for neurosurgery, it still has some limitations. As a single center study, the number of samples are relatively small and the causes of diseases are few. Variables were not fully included into to analysis due to data availability. Multi-center studies with larger sample size and more indices are needed to further validate our findings.