Discussion On The Influencing Factors Of Acute Altitude Sickness And Preliminary Establishment Of Classification Model

Objective When individuals who reside in plain go to high altitude or who reside in plateau go to higher altitude,a series of symptoms will occur as a result of acute hypoxia,which is also called acute mountain sickness(AMS).It can sometimes progress to high-altitude pulmonary edema(HAPE)and high-altitude cerebral edema(HACE),which would be life threatening.Due to more hypoxic and colder environment,the incidence of AMS of Chinese Antarctic Kunlun Station which is located at the South Pole Dome A region(4087m)is much higher than that of other plateaus with the same altitude.In order to minimize the risk of severe AMS during expedition,pre-expeditioners would go to Tibet to be screened before departure to Antarctica.This research aims to screen predisposing factors and establish classification models by analyzing baseline parameters,which provide references for selecting eligible Antarctica expeditioners.Methods 1.A total of 149 samples in database of the inland pre-selected corhorts were included.122 samples were randomly selected as training dataset,and the remaining 27 samples were used as test dataset.Comparisons between two groups and univariate logistic analysis were conducted by IBM SPSS software.The logistic regression(LR),decision tree(DT)model and random forest(RF)model were constructed by open source R software.Model evaluation parameters included accuracy,sensitivity,specificity and area under curve(AUC).After that,the importance of each variable was calculated and ranked.2.The effect of overweight and obesity on the occurrence of AMS was analyzed.Literatures which studied association between overweight,obesity and AMS during 2007-2016.RevMan 5.3 software was utilized for Meta analysis.Results 1.There were significant differences in pulse oxygen saturation,basic information,cardiovascular and pulmonary function,mental status between AMS group and non-AMS group.Based on comparisons between groups and univariate analysis results,age,overweight and obesity,ascent profile(plane or train),thoracic fluid content(TFC),systolic time ratio(STR),systemic vascular resistance index(SVRI)and total mood disturbance(TMD)were included.All models were effective in classifying the AMS and non-AMS group.In the training dataset,the accuracy,sensitivity,specificity and AUC of the logistic model were 0.738,0.685,0.816 and 0.787,respectively.TFC,STR TMD and ascent profile were significant determinants that were correlated with the occurrence of AMS.For decision tree,these parameters were 0.803,0.849,0.735 and 0.833,respectively.TFC,STR,age and TMD were important split nodes.In the random forest model,these parameters were 1.000,1.000,1.000 and 1.000,respectively.Variable importance analysis showed that STR,TFC,TMD and ascent profile were important factors that were related with AMS.2.For Meta analysis,37 studies were primarily selected.According to inclusive and exclusive criteria,6 studies and 6323 cases were finally included for meta-analysis in relationship between overweight,obesity and AMS.The results showed that the risk of AMS in overweight and obese group significantly increased compared with those in the control group,and the pooled odds ratio(OR)was 2.39(95%CI[1.21,4.70]).Conclusion All three models achieved good performance in classifying and predicting the occurrence of AMS.STR,TFC,TMD,ascent profile and overweight and obesity were important factors that were associated with occurrence of AMS.Our research provides references for screening AMS susceptible expeditioners.