| ObjectiveAcute pancreatitis(AP)is an acute systemic inflammation of variety of causes,it is a common disease and frequently encountered disease in the clinical.In recent years,its incidence has been increasing year by year and the fatality rate remains high,especially severe acute pancreatitis.Multiple scoring systems and serological tests have been used to assess and predict the severity of acute pancreatitis,but for various reasons,they are limited in clinical practice.Therefore,the purpose of this study aims to explore the value of NLR in predicting the severity of acute pancreatitis,so that to find more clinical value of serological diagnosis model,contributing to the follow-up treatment and the improvement of prognosis.MethodIn this study,240 cases of acute pancreatitis patients from August 2014 to October 2017 were collected using a retrospective study.Including general information(gender,age,etiology)and laboratory data(white blood cell count,absolute value of neutrophil,absolute value of lymphocyte,blood calcium,triglycerides,creatinine,urea,liver lactate dehydrogenase),and calculating the ratio ofneutrophil and lymphocyte namely NLR values.The 240 patients with acute pancreatitis were classified into mild and severe cases according to the classification criteria for the severity of acute pancreatitis in Atlanta in 2012.According to the preference score matching,the nearest neighbor matching method was matched,and finally 43 pairs were included in the study.Kolmogorov-smirnov test was used to test the normality of data.The mean plus or minus the standard deviation was used for measurement data according with normal distribution.Median(interquartile range)was used for measurement data of no according with normal distribution.frequency and rate(%)or the proportion(%)was used for the count data.In the measurement data,if the data was normal distribution and the variance was homogeneous between the two groups,t-test was selected.If the data was not normal distribution,the mann-whitney U test was selected.The comparison between two or more constituent ratios was based on the chi-square test or Fisher's exact probability method.The difference of NLR in the two groups and the change trend of NLR with admission time were analyzed by repeated measurement variance.Spearman correlation was used to analyze the correlation between severity of disease and NLR.The accuracy of the diagnosis of disease severity evaluation in different time of admission was analyzed using the area under the ROC curve,sensitive,specific degree index.As different admission time,The NLR comparison of area under the ROC curve used Stata14 software for analysis.All the above results were double-sided,and the difference was statistically significant with P<0.05.ResultIn front of the propensity score matching,age(z=1.135,P=0.256),gender(?~2=1.035,P=0.361)and white blood cell count(z=0.350,P=0.726)were no statistically between mild and severe acute pancreatitis patients.two groups in triglycerides(z=3.900,P<0.001),blood calcium(z=7.009,P<0.001),urea(z=3.912,P<0.001),creatinine(z=3.881,P<0.001),liver lactate dehydrogenase(z=6.708,P<0.001),and etiological distribution(Fisher's P=0.002)were statistically significant.The blood calcium of the patients in the severe group was significantly lower than that in the mild group.Triglycerides,urea,creatinine,liver lactate dehydrogenase were significantly higher than those in the mild group.The etiology of mild acute pancreatitis was mainly biliary,and severe acute pancreatitis was characterized by hypertriglyceridemia.After propensity score matching,mild and severe acute pancreatitis patients in age(z=1.151,P=0.880),gender(?~2=0.508,P=0.635),white blood cell count(z=0.354,P=0.723),triglycerides(z=0.505,P=0.613),blood calcium(t=0.414,P=0.680),urea(z=1.304,P=0.192),creatinine(z=0.946,P=0.344),lactate dehydrogenase(z=0.833,P=0.405),and etiological distribution(Fisher's P=0.934)were no statistical differences.According to the results of Repeated measures design analysis,the NLR between mild and severe acute pancreatitis patients were statistically significant(F=10.063,P=0.002),The NLR of different time on admission and disease severity were no interaction between(F=0.199,P=0.749).The NLR between mild and severe acute pancreatitis patients on admission,24 hours after admission and 48hours after admission were statistically significant(F=10.695,P<0.001).Compared two(Sidak method)the results show that,on admission and 24 hours after admission,NLR were statistically significant(P=0.025),on admission and 48 hours after admission,NLR were statistically significant(P<0.001).On 24 hours and 48 hours after admission,NLR were not statistically significant(P=0.068).The change trend of NLR in the patients with mild and severe group was in linear decline during admission,24 hours after admission and 48 hours after admission(F=15.712,P<0.001).The Spearman correlation analysis was used between NLR on admission,24hours after admission and 48 hours after admission and the illness severity of acute pancreatitis.The results showed the NLR of three point on admission were positively correlated with severity of AP,and the correlation was 48h after admission NLR>on admission NLR>24h after admission.The ROC curve was drawn and the area AUC was calculated according to the diagnostic accuracy of NLR at different time of admission to the severity of the disease.The AUC of NLR diagnosed with severe acute pancreatitis on admission was0.660(95%confidence interval 0.544 to 0.776).The AUC of NLR diagnosed with severe acute pancreatitis on 24h after admission was 0.654(95%confidence interval was 0.538 to 0.770).The AUC of 48h after admission NLR for the diagnosis of severe acute pancreatitis was 0.746(95%confidence interval was 0.641 to0.850.Stata14 was used to test the area under the ROC curve of NLR,24hNLR and48h NLR after admission to the hospital,and it was found that the difference was not statistically significant(?~2=5.970,P=0.051).The diagnostic truncation value of NLR was determined according to the maximum method of Youden index.The truncation value of NLR on admission was14.76,the sensitivity was 60.5%,and the specificity was 72.1%.On 24 hours after admission,the truncation value was 10.49,the sensitivity was 67.4%,and the specificity was 65.1%.The truncation value of 48 hours after admission NLR was6.61,the sensitivity was 93.0%,and the specificity was 53.5%.The related sample inspection Cochran Q was used for the test of NLR sensitivity and specific degrees on admission,24h and 48h after admission.In the three indexes,sensitivity was statistically significant difference(Cochran Q statistic=15.524,P<0.001).Compared two(Bonferroni method)the results showed the sensitivity difference between 48 hours after admission NLR and on admission NLR was statistically significant(P=0.001).24 hours and 48 hours after admission,the sensitivity difference of NLR was statistically significant(P=0.010).NLR on admission and 24 hours after admission was no difference(P=1.000).The specific degrees of the three indicators was statistically significant(Cochran Q statistic=7.538,P=0.023).compared two(Bonferroni method)the results showed that the specific degree difference of NLR was statistically significant(P=0.020)on admission and 48 hours after admission.The specific degree difference of NLR with24 hours and 48 hours after admission were not statistically significant(P=0.268).On admission and 24 hours after admission,NLR were no difference(P=0.925).Conclusion1.Biliary factor is still the main cause in mild acute pancreatitis,but the main etiology of severe acute pancreatitis is hypertriglyceridemia.2.NLR is positively correlated with the severity of acute pancreatitis,and the higher the NLR,the more severe acute pancreatitis.3.The truncation value of NLR in different time points after admission is different and dynamiac.4.Dynamic monitoring of NLR level in acute pancreatitis patients can be regarded as one of the indicators of severity evaluation.5.The NLR level on 48h after admission can better reflect the severity of disease. |
PDF Full Text Request