Risk factors of pulmonary complications after spine surgery and a risk assessment table was established: a retrospective cohort study

Purpose To determine the risk factors for pulmonary complications after spine surgery and establish a risk assessment table. Methods A total of 627 were retrospectively collected and studied. Univariate analysis revealed signicant variables. Risk factors and the effects of the model were determined by unconditional binary logistic regression analyses and the receiver operating characteristic (ROC) curve, respectively. A risk assessment scale for patients after spine surgery was also established based on the risk factors.


Introduction
Spinal diseases are some of the most frequently encountered problems that affecting people's quality of life in particular [1] . It was reported that the incidence and prevalence of myelopathy due to degeneration of the spine were estimated at a minimum of 41% and 605 per million in North America respectively, and surgical rates caused by the disease was also on the rise [2] . Although most spinal diseases are not fatal, their incidence rate has caused huge losses from the individual and social perspective [3] . Injury complexity and propensity for related complications would happen because of the spinal column's intimate association with multiple vital structures [4] . Complications are the most concerned problems of patients and physicians, once they occur, they may have personal and economic consequences, and affect the quality of life and future independence of patients [5] .
Prior studies have reported complications after spine surgery [6,7] . Notably, pulmonary complications are especially responsible for morbidity and mortality, as well as prolonged hospitalization, generate substantial economic burden, and compromised quality of life [8,9] . Risk factors for pulmonary complications after spine surgery had been identi ed in much research [10][11][12][13] , and its rates vary from study to study, reported to range from 0.9-9% [7,10,14,15] . At the same time, there is no report about establishing a risk assessment table. The identi cation and quanti cation of risk factors for postoperative complications of spinal surgery are very important for both patients and clinicians. In addition to the obvious importance to patient safety, risk factor information becomes critical as health care policy makers implement "quality" indicators. [16] Age and surgical invasiveness had been reported as signi cant risk factors for patients who underwent spine procedures to infect complications during perioperative period [17] . It was also reported that based on these risk factors clinicians could stratify the risk of perioperative complications in patients undergoing spinal surgery [17,18] and these results could be used to establish prediction models whereby the probability of a postoperative complication can be predicted for each patient who undergoes spine surgery.
Therefore, patients who underwent spine surgery at Hebei General Hospital from January 2018 to December 2019 were included in this study. Risk factors of pulmonary complications after spine surgery would been identi ed. Meanwhile, we attempted to establish a risk assessment scale for such patients based on the relevant literatures and our ndings, which may be e cient for doctors and nurses to assess the risk of postoperative pulmonary complications.

Patients
Patients who underwent spine surgery at Hebei General Hospital from January 2018 to December 2019 were retrospectively collected and studied. Inclusion criteria of patients were as follows: 1) age ≥ 18 years; 2) ful lled the follow-up no less than six months after surgery; 3) diagnosed by spinal surgery in hospital. The exclusion criteria of patients were as follows: 1) with severe pulmonary disease before operation; 2) were diagnosed with combining surgeries in other parts; 3) had incomplete data; 4) age < 18 years.

Diagnostic of pulmonary complication
Postoperative pulmonary complications (PPCs) could be broadly de ned as conditions affecting the respiratory tract that can adversely in uence the clinical course of a patient after surgery [19,20] .
Nevertheless, PPCs de nitions usually include pneumonia, exacerbation of chronic obstructive pulmonary disease, respiratory failure, bronchospasm, pneumothorax, pleural effusion, atelectasis, and various forms of upper airway obstruction.
In this study, fever, pulmonary in ammation, respiratory failure, atelectasis, respiratory tract infection, pleural effusion, venous thrombosis of lower extremity, urinary infection, arrhythmia,gastrointestinal discomfort and so on were all the complications after operation. Pulmonary complication was diagnosed by patients' symptoms of pulmonary in ammation, upper respiratory tract infection, pulmonary embolism and pleural effusion.

Risk factors
Gender, age, BMI, smoking, alcohol abuse, hypertension, diabetes, prior spine surgery, type of spinal diseases, type of operation, surgical site, operation time limit, operation start time, type of anesthesia, ASA scores, length of stay in operating room, duration of operation, preoperative hospitalization days and hospitalization days were used to analyze the risk factors of pulmonary complications after spine surgery.

Statistical Analysis
Statistical analysis was performed using SPSS 21.0 (IBM). Continuous variables were expressed as means ± standard deviation and analyzed by independent samples T test. χ 2 test was used for categorical and discrete variables. Meaningful factors and risk factors considered by clinical experts were added to unconditional binary logistic regression. Risk factors were searched by unconditional binary logistic regression, which were independent variables with univariate analysis of P < 0.05, and the dependent variables were with or without pulmonary complication. The ROC curve was used to evaluate the effect of the logistic regression model. P value of less than 0.05 was considered signi cant.

Establishment of the risk assessment scale
A risk assessment scale for patients after spine surgery was established based on the results of logistic regression analysis. The independent risk factors were grouped according to clinical signi cance or usage habits. For each risk factor, an appropriate group was selected as the basic risk reference value, and the score of this group was recorded as 0 in the subsequent construction of the rating scale. In this study, the β value of age was used as the reference value, and each component value was calculated according to the β value and strati cation of each risk factor. According to the opinion of clinical experts, other risk factors (hypertension, chronic pulmonary diseases, preoperative LOS, duration of operation) were added to the risk assessment table too. Adjust each evaluation factor reasonably and construct a risk assessment scale.

Results
A total of 656 patients were enrolled in this study. However, among them, 8 patients with severe pulmonary disease before operation, 4 patients were younger than 18 years old and 17 patients without following up were excluded. Finally, a total of 627 patients were included in this retrospective study, among which 49 patients (7.81 %) had pulmonary complications after surgery (Table 1).  Table 2). Meaningful factors and risk factors (prior spine surgery, type of anesthesia, duration of operation) considered by clinical experts were added to unconditional binary logistic regression analysis. The variable assignment of each risk factor was shown in Table 3.   Area under the receiver operating characteristic curve (AUC) can be used to measure the reliability of the model, with values close to 1.0 indicating high diagnostic accuracy [21] . AUC was 0.883 (95% CI: 0.843-0.923, Fig. 1 and Table 5), which meant the above indicators were reliability.  There were 13 evaluation indicators including preoperative and postoperative evaluation indicators. Age, BMI, smoking, hypertension, diabetes, chronic pulmonary diseases, type of spinal diseases, surgical site, were preoperative evaluation indicators. The postoperative evaluation indicators include preoperative hospitalization days, ASA scores, type of anesthesia, type of operation and duration of operation were postoperative evaluation indicators. The highest assessment score of preoperative is 21, postoperative is 18, and the total is 39 points of all these indicators ( Table 7). The higher the score, the higher the risk of pulmonary complications after spine surgery.

Discussion
It is estimated that more than 230 million major operations occur annually in the world [22] . Surgeons and patients are very concerned about the possibility and likelihood of a postoperative medical complication after surgery, which will affect their decision-making. Prior to extensive spinal surgery, high-risk patients are often evaluated by medical providers for risk strati cation and health optimization [23] . Spinal surgery is no exception. PPCs are common, and can be considered as a composite outcome measure [24,25] .
PPCs had been shown to be more common than cardiac complications in some studies [26,27] . Although many scoring systems can be used to quantify PPCs risk, there is no consensus on the best one to use, especial for spine surgery, and they remain too complex to use clinically [24] .
In the present study, we retrospectively evaluated patients who underwent spine surgery at our hospital from January 2018 to December 2019. points of all these indicators. The higher the score, the higher the risk of pulmonary complications after spine surgery.
Obesity has been illustrated to reduce lung volume and change ventilation pattern [28] , which is more likely to occur in the elderly patients. In addition, obesity was revealed to be a risk factor in certain studies, which examined the prevalence and risk factors of pulmonary embolism (PE) after spinal surgery [29] . Patients with diabetes, two or more spinal diseases are also at a signi cantly increased risk for complication. Compared with other parts of the spine, thoracic surgery is more likely to affect the lung function. Many studies also show that minimally invasive surgery has the advantages of small incision, small trauma, fast recovery, short hospital stay, and can effectively reduce the surgical complications, including pulmonary complications. The higher the anesthesia score, the more serious the patient's condition is. In addition, because of the combined intravenous and respiratory anesthesia, it is bound to affect the function of the lung. These risk factors are consistent with previous reports [10] .
This study has several limitations. Firstly, the data were retrospectively collected in a single center which might lead to an unavoidable selection bias. Secondly, the established risk assessment table in our study has not been used in clinic. In the next study, we will conduct prospective studies to verify the validity of this risk assessment table. Furthermore, the sample sizes in this study are relatively small, and further study with a larger number of samples is warranted. These limitations notwithstanding, the risk factors of pulmonary complications after spine surgery were determined in our study. Furthermore, the risk assessment table was established, which could assess the risk of postoperative pulmonary complications.
In conclusion, targeted prevention interventions should be used to prevent pulmonary complications after spine surgery, in order to comprehensively manage risk factors in the perioperative period.  Area under the curve of the predicted meaningful and risk factors.