Analysis Of Influencing Factors Of Diabetes Insipidus After Pituitary Adenoma Endoscopic Transsphenoidal Surgery And Establishment Of Prediction Model

ObjectiveTo summarize the medical records of 82 patients who underwent endoscopic transsphenoidal resection of pituitary adenomas in the First Affiliated Hospital of Xiamen University from 2009 to 2019,analyze the possible influencing factors of postoperative diabetes insipidus of pituitary adenomas,and establish the predictive model of postoperative diabetes insipidus.MethodSPSS 22.0 software was used for statistical analysis.The measurement data that meet the normal distribution adopt mean±standard deviation.The classification data were expressed by percentage,and the comparison between the two groups was performed by chi square test.The difference was statistically significant when P≤0.05.The independent risk factors were introduced into R software(R.4.0.2),and the nomogram prediction model was established.C-index was used to evaluate the discrimination of nomogram model.The range of C-index was 0.5-0.7 with low discrimination;0.71-0.9 with medium discrimination;greater than 0.9 with high discrimination.Calibration chart is used to evaluate the consistency between the predicted risk and the actual risk.ResultsAmong 82 cases of pituitary adenoam surgery,20 cases had diabetes insipidus after surgery,and the incidence of diabetes insipidus was 24.3%.Statistical analysis showed that tumor volume was an independent influencing factor of postoperative diabetes insipidus(W=11.277,P=0.001,OR=1.530).Cavernous sinus invasion was an independent risk factor for postoperative diabetes insipidus(W=4.968,P=0.026,OR=4.950).Deformation degree of the third ventricle was an independent risk factor for diabetes insipidus(W=5.307,P=0.021,OR=0.065).Objective to establish a nomogram model of diabetes insipidus risk.Deformation degree of the third ventricle was 0 score,no severe deformation was 19 score.Cavernous sinus invasion was 20 score,no cavernous sinus invasion was 0 score.Volume<5cm~3,0score;volume>=5cm~3and<=9cm~3,2 score;volume>=10cm~3and<=14cm~3,18 score;volume>=15cm~3and<=19cm~3,48 score;volume>=20cm~3,100score.The C-index of nomogram model for predicting postoperative diabetes insipidus was 0.83,with good discrimination.Calibration evaluation found that the prediction line was close to the standard line,and the calibration ability was good,which indicated that the prediction risk of the model was consistent with the actual risk.Conclusion1.The incidence of diabetes insipidus after pituitary adenoma surgery is related to tumor size,cavernous sinus invasion and deformation degree of the third ventricle,but not to age,gender,BMI,endocrine function type,surgical approach and tumor resection degree.The volume of pituitary adenoma larger than 10.15cm~3had a higher incidence of diabetes insipidus.2.The C-index of postoperative diabetes insipidus predicted by the line graph model was 0.83.It is considered to be a good discrimination.Calibration evaluation found that the prediction line was close to the standard line,and the calibration ability was good,which indicated that the prediction risk of the model was consistent with the actual risk.