| All parts of body can be injuried by blunt or penetrating trauma which can be traffic accidents or other ones,with the development of social economy and the change of life style.The diaphragm worked as an anatomic barrier between the thoracic and abdominal cavities is seldom present as an isolated injury because of its special site of body,which always associates with mulitiple injuries to some extent.Grimes divided this special disease into 3 phases according to its natural evolution process,which are the initial or acute phase,latent phase,and obstructive phase.(1) the acute phase,which encompasses the period from the trauma to recovery from the primary lesions and is usually dominated by the presence of associated lesions other thanDI or TDH itself;(2)the latent phase,which is related to visceral herniation into the thorax and is usually manifested by chest or abdominal pain and gastrointestinal symptoms;and(3) the obstructive phase,characterized by signs of obstruction or ischemia of the viscera trapped in the diaphragm defect.According to the view of Grimes.DI refers to the rupture of the diaphragm itself,while TDH refers to herniation of abdominal viscera into the thorax on the basis of DI.We describe the 3 phases of Grimes as two stages of this disease,which are DI stage and TDH stage.Late or missing diagnosis can be seen very often as result of the preoccupied thoughts of the surgeon because of some life-threating asscociated injuries like shock, coma and profuse haemorrhage and the poor understanding of the characteristics of the regularity of this special trauma.At present,almost all experts agree that keeping a high index of suspision is most important to the diagnosis of DI and TDH,however, the primary and essential conditon of the formation of such vigilant psychology are the understanding of the mechanism of this special trauma and the deep apperception of the essence of this disease.A retrospective study was performed for the clinic data from the 51 patients admitted with DI or TDH in our department from 1993～2008 in order to depict the characteristics of the regularity of this disease,which can be applied to the future clinic practice.From the literature review for DI and TDH,we discover that although some factors can influence the prognosis of them significantly,which includes age,sex,ISS (Injuries Severity Score),volume of transfusion,the type of trauma,the side of ruptured diaphragm,the stage of disease progress,the time of diagnosis,shock symptoms when admitted to hospital,after-operative severe complications and sever associated injuries,different opinions and diaputes among many studies about these factors have not reached consensus yet.At present there is no systematic review of the quantitative analysis for such clinic studies aimed at such clinic problem in China and abroad.For this reason,we develop a Meta-analysis of some possible factors influencing mortality of DI and TDH to identify risk factors and the extent of risks.The diagnostic accuracy of DI an TDH preoperatively is low from the studies between China and abroad.At the area of the imaging,the ability of the conventional X-ray is insufficient obviously.Along with the high resolution CT applied to the clinic for diagnsising DI and TDH,the accuracy estimates can be seen from various studies.At present,most studies agree that CT has a good performance for diagnosis, however,the discrepancy of the accuracy estimates and some relatve small sample size make the evaluation of CT performance with some discount.For the same reason, we collect some relative studies to systematically assess the diagnostic performance of CT for DI and TDH.At the condition of different imaging sign,different type of injury and different side of ruptured diaphragm,we develop a Meta-analysis for evaluating these discrepancies.Methods:The first section of this study described the background and aim and pointed out the significance and method of the study.In the second secton of this study,a retrospective study was performed for the clinic data from the 51 patients admitted with DI or TDH in our department from 1993～2008 in order to depict the characteristics of the regularity of this disease and to identify the associated factors influencing preoperative diagnosis and mortality, which can be applied to the future clinic practice.In the third section,we developed a Meta-analysis of some possible factors influencing mortality of DI and TDH,which includes the type of trauma,sex,the side of ruptured diaphragm,the stage of disease progress,the time of diagnosis,shock symptoms when admitted to hospital,after-operative severe complications and sever associated injuries.In the forth section,a diagnostic test Meta-analysis was developed to assess the diagnostic performance of CT for DI and TDH systematically.At the condition of different imaging sign,different type of injury and different side of ruptured diaphragm,we develop a Meta-analysis for evaluating these discrepancies. Results:First,results of the second section:1,There were 25 cases with blunt trauma and 26 with penetrating trauma.When compared with penetrating trauma group,the patients of blunt group were characterized by the elder age,the higher ratio of TDH stage of this disease,the higher ratio of the herniated number beyond 1,the longer diameter of the ruptured diaphragm,and the higher ratio of preoperative diagnostic number.2,The sensitivity of X-ray when diagnosing DI and TDH was 48.3%,with 40%for CT.The ealy diagnostic ratio was 86.3%.The ratio of X-ray was 22.7%among the relative way of early diagnsis,with exploration of laporotomy or thoracotomy being 72.7%.3,The ratio of the preoperative diagnosis was 29.4%and the ratio of X-ray was 86.7%among the relative way of preoperative diagnsis.Univariate analysis identified 9 factors were different for preoperative diagnosis singnifantly,which were ISS,size of ruptured diaphragm,prehospital time,time before surgery after hospital,type of injury(blunt or penetrating),side of injuried diaphragm(left or right),shock symptoms when admitted to hospital(whether or not),preoperative X examination,different stage of the disease(DI or TDH). Multivariate Logistic stepwise regression demonstrated that repeat X examination preoperatively and the blunt trauma were two independent predictors for increasing the rate of preoperative diagnosis.4,9 patients died and overall mortality was 17.6%.4 died of hemorrhagic shock at early stage(＜24 h)and 5 died of postoperative sever complications at late stage (＞24 h),including sever infections and ARDS(n=2),sever infections(n=1), ARDS(n=1) and MODS(n=1).7 factors identified by univariate analysis were different statistically for prognosis,which were ISS＞22,age＞40,volume of transfusion≥1400ml,shock symptoms when admitted to hospital,after-operative severe complications,systolic blood pressure when admitted to hospital≤80mmHg,and splenic injury scale≥Ⅲ.Multivariate Logistic stepwise regression demonstrated that Volume of transfusion≥1400ml and spleen injury scale≥Ⅲwere both the independent indicators for mortality.Second,results of the third section:150 studies were searched out according to the criterion of including and excluding,among of which 17 studies were adopted for this analysis.There were 1535 patients enrolled and 8 factors analyzied for Meta-analysis,which the type of trauma,sex,the side of ruptured diaphragm,the stage of disease progress,the time of diagnosis,shock symptoms when admitted to hospital,after-operative severe complications and sever associated injuries.1,Sever associated injuries on prognosis of patients with DI and TDH1224 patients were enrolled from 11 retrospective studies,among of which 1083 patients were calculated in sever AI(associated injuries) group and 141 in non-sever AI group.There were 252 patients died in sever AI group and 9 in non-sever AI group,with the mean mortality rate was 21.3%.Meta-analysis demonstrated that the pooled OR of sever AI on mortality was3.36(95%CI:1.81～6.24,P=0.0001) when compared with non-sever AI.The test of heterogeneity showed that there was no statistical heterogeneity among studies included in the Meta-analysis(P=0.69),which indicated that fixed effect model was suitable for the synthesis.The funnel plot was asymmetry which indicated the exiting of potential publication bias.2,Type of trauma on prognosis of patients with DI and TDH. 1194 patients were enrolled from 9 retrospective studies,among of which 852 patients were calculated in penetrating group and 342 in blunt group.There were 175 patients died in penetrating group and 81 in blunt group,with the mean mortality rate was 21.4%.Meta-analysis demonstrated that the pooled OR of penetrating trauma on mortality was 0.62(95%CI:0.27～1.41,P=0.25) when compared with blunt trauma. The test of heterogeneity showed that there was statistical heterogeneity among studies included in the Meta-analysis(P=0.009),which indicated that random effect model was suitable for the synthesis.The funnel plot was basically symmetry which indicated no publication bias.3,Side of ruptured dipharagm on prognosis of patients with DI and TDH.949 patients were enrolled from 7 retrospective studies,among of which 619 patients were calculated in left-sided group and 330 in right-sided group.There were 105 patients died in left-sided group and 68 in right-sided group,with the mean mortality rate was 18.2%.Meta-analysis demonstrated that the pooled OR of left-sided trauma on mortality was 0.76(95%CI:0.54～1.07,P=0.11) when compared with right-sided trauma.The test of heterogeneity showed that there was no statistical heterogeneity among studies included in the Meta-analysis(P=0.39),which indicated that fixed effect model was suitable for the synthesis.The funnel plot was asymmetry which indicated the exiting of potential publication bias.4,Stage of the disease on prognosis of patients with DI and TDH.190 patients were enrolled from 3 retrospective studies,among of which 63 patients were calculated in DI group and 127 in TDH group.There were 12 patients died in DI group and 25 in TDH group,with the mean mortality rate was 19.5%.Meta-analysis demonstrated that the pooled OR of DI stage on mortality was 0.98(95%CI:0.46～1.10,P=0.98) when compared with TDH stage.The test of heterogeneity showed that there was no statistical heterogeneity among studies included in the Meta-analysis(P=0.82),which indicated that fixed effect model was suitable for the synthesis.5,Sever postoperative complications on prognosis of patients with DI and TDH.369 patients were enrolled from 7 retrospective studies,among of which 159 patients were calculated in sever postoperative complication(POC) group and 210 in non-sever POC group.There were 28 patients died in sever POC group and 5 in non-sever POC group,with the mean mortality rate was 8.9%.Meta-analysis demonstrated that the pooled OR of sever POC on mortality was 5.23(95%CI:2.22～12.28,P=0.0001) when compared with non-sever POC.The test of heterogeneity showed that there was no statistical heterogeneity among studies included in the Meta-analysis(P=0.28),which indicated that fixed effect model was suitable for the synthesis.The funnel plot was basically symmetry which indicated no publication bias.6,Sex on prognosis of patients with DI and TDH.232 patients were enrolled from 5 retrospective studies,among of which 176 patients were calculated in male group and 210 in female group.There were 33 patients died in male group and 15 in female group,with the mean mortality rate was 20.7%.Meta-analysis demonstrated that the pooled OR of male on mortality was 0.66 (95%CI:0.33～1.33,P=0.26) when compared with femal.The test of heterogeneity showed that there was no statistical heterogeneity among studies included in the Meta-analysis(P=0.91),which indicated that fixed effect model was suitable for the synthesis.The funnel plot was basically symmetry which indicated no publication bias.7,Time of diagnosis on prognosis of patients with DI and TDH. 224 patients were enrolled from 5 retrospective studies,among of which 199 patients were calculated in ED(early diagnosis) group and 25 in LD(late diagnosis) group. There were 33 patients died in ED group and 1 in LD group,with the mean mortality rate was 15.2%.Meta-analysis demonstrated that the pooled OR of ED on mortality was 1.87 (95%CI:0.54～6.48,P=0.32) when compared with LD.The test of heterogeneity showed that there was no statistical heterogeneity among studies included in the Meta-analysis(P=0.86),which indicated that fixed effect model was suitable for the synthesis.The funnel plot was asymmetry which indicated the exiting of potential publication bias.8,Shock symptoms when admitted to hospital on prognosis of patients with DI and TDH.215 patients were enrolled from 4 retrospective studies,among of which 64 patients were calculated in shock symptoms group and 151 in non-shock symptoms group.There were 26 patients died in shock symptoms group and 10 in non-shock symptoms group,with the mean mortality rate was 16.7%.Meta-analysis demonstrated that the pooled OR of shock symptoms on mortality was 13.00(95%CI:3.25～52.02,P=0.0003) when compared with non-shock symptoms.The test of heterogeneity showed that there was no statistical heterogeneity among studies included in the Meta-analysis(P=0.16),and random effect model was also suitable for the synthesis here.Third,results of the fourth section:Results of the first segment1191 studies were searched and 8 studies were selected,with 728 patients involved.The pooled accuracy estimates for the 8 studies like DOR(Diagnostic Odds Ratio),sensitivity and specificity were 45.74(95%CI 23.39～89.44),82% (95%CI:0.78～0.86) and 92%(95%CI:0.89～0.95) respectively,withâ… ~2 of heterogeneity test being 22.1%,32.5%and 82%,and random effect model was also suitable for the synthesis here.Results of the second segment1,The diagnostic value for diaphragmatic defect signMeta-analysis demonstrated that the pooled OR of diaphragmatic defect sign was 30.51(95%CI:13.86～67.17,P=0.00001).The test of heterogeneity showed that there was no statistical heterogeneity among studies included in the Meta-analysis(P=0.77),which indicated that fixed effect model was suitable for the synthesis.The funnel plot was asymmetry which indicated the exiting of potential publication bias.Meta-analysis of diagnostic test demonstrated that the pooled sensitivity and specificity were 64%(95%CI:0.53～0.73)and 95%(95%CI:0.91～0.97).The test of heterogeneity for sensitivity and specificity showed that there were statistical heterogeneity among studies included in the Meta-analysis(â… ~2=79%and 41.5%), which indicated that random effect model was suitable for the synthesis.2,The diagnostic value for intrathoracic visceral herniation sign.Meta-analysis demonstrated that the pooled OR of intrathoracic visceral herniation sign was 65.07(95%CI:8.43～502.36,P=0.0001).The test of heterogeneity showed that there was statistical heterogeneity among studies included in the Meta-analysis(P=0.01),which indicated that random effect model was suitable for the synthesis.Meta-analysis of diagnostic test demonstrated that the pooled sensitivity and specificity were 60%(95%CI:0.50～0.71)and 98%(95%CI:0.95～0.98).The test of heterogeneity for sensitivity and specificity showed that there were statistical heterogeneity among studies included in the Meta-analysis(â… ~2=92.7%and 70.6%), which indicated that random effect model was suitable for the synthesis.3,The diagnostic value for collar sign.Meta-analysis demonstrated that the pooled OR of collar sign was 41.95(95%CI: 10.99～160.11,P=0.00001).The test of heterogeneity showed that there was no statistical heterogeneity among studies included in the Meta-analysis(P=0.40), which indicated that fixed effect model was suitable for the synthesis.The funnel plot was asymmetry which indicated the exiting of potential publication bias.Meta-analysis of diagnostic test demonstrated that the pooled sensitivity and specificity were 42%(95%CI:0.31～0.53) and 100%(95%CI:0.99～1.00).The test of heterogeneity for sensitivity showed that there was statistical heterogeneity among studies included in the Meta-analysis(â… ~2=73.4%),which indicated that random effect model was suitable for the synthesis.4,The diagnostic value for dependent viscera sign.Meta-analysis demonstrated that the pooled OR of dependent viscera sign was 42.03(95%CI:0.97～1826.57,P=0.05).The test of heterogeneity showed that there was statistical heterogeneity among studies included in the Meta-analysis (P＜0.00001),which indicated that random effect model was suitable for the synthesis.Meta-analysis of diagnostic test demonstrated that the pooled sensitivity and specificity were 60%(95%CI:0.47～0.72) and 94%(95%CI:0.91～0.97).The test of heterogeneity for sensitivity and specificity showed that there were statistical heterogeneity among studies included in the Meta-analysis(â… ~2=43.9%and 95.8%), which indicated that random effect model was suitable for the synthesis.5,The diagnostic value for segmental diaphragmatic thickening sign. Meta-analysis demonstrated that the pooled OR of segmental diaphragmatic thickening sign was 12.42(95%CI:1.42～108.95,P=0.02).The test of heterogeneity showed that there was statistical heterogeneity among studies included in the Meta-analysis(P=0.005),which indicated that random effect model was suitable for the synthesis.Meta-analysis of diagnostic test demonstrated that the pooled sensitivity and specificity were 46%(95%CI:0.32～0.61)and 82%(95%CI:0.90～0.97).The test of heterogeneity for specificity showed that there was statistical heterogeneity among studies included in the Meta-analysis(â… ~2=82%),which indicated that random effect model was suitable for the synthesis.6,The comparison of CT performance for diagnosing diaphragma left-sided and right-sided injuries.Meta-analysis demonstrated that the pooled OR of CT for diagnosing diaphragma left-sided injuries was 125.67(95%CI:46.55～339.28,P＜0.00001) and for diaphragma right-sided injuries was 27.02(95%CI:8.65～84.39, P＜0.00001).The test of heterogeneity for them showed that there were no statistical heterogeneity among studies included in the Meta-analysis(P=0.84 and 0.23), which indicated that fixed effect model was suitable for the synthesis.Meta-analysis of diagnostic test demonstrated that the pooled sensitivity and specificity of CT for diagnosing diaphragma left-sided injuries were 97%(95%CI: 0.94～0.99) and 96%(95%CI:0.94～0.98) and for diagnosing diaphragma right-sided injuries were 74%(95%CI:0.57～0.88) and 97%(95%CI:0.94～0.99). The test of heterogeneity for specificity of CT for diagnosing diaphragma left-sided injuries and for right-sided injuries showed that there were statistical heterogeneity among studies included in the Meta-analysis(â… ~2=49.8%and 68.6%),which indicated that random effect model was suitable for the synthesis. Meta-analysis demonstrated that the pooled OR of CT for diagnosing diaphragma left-sided and right-sided injuries was 3.27(95%CI:0.96～11.23, P=0.06).The test of heterogeneity for them showed that there were statistical heterogeneity among studies included in the Meta-analysis(P=0.32),which indicated that fixed effect model was suitable for the synthesis.The funnel plot was asymmetry which indicated the exiting of potential publication bias.7,The comparison of CT performance for blunt and penetrating injuries.Meta-analysis demonstrated that the pooled OR of CT for diagnosing blunt injuries was 68.69(95%CI:25.24～186.97,P＜0.00001) and for penetrating injuries was 77.27(95%CI:31.29～190.82,P＜0.00001).The test of heterogeneity for them showed that there were no statistical heterogeneity among studies included in the Meta-analysis(P=0.34 and 0.29),which indicated that fixed effect model was suitable for the synthesis.Meta-analysis of diagnostic test demonstrated that the pooled sensitivity and specificity of CT for blunt injuries were 80%(95%CI:0.70～0.88) and 95%(95%CI: 0.92～0.97) and for penetrating injuries were 85%(95%CI:0.77～0.91) and 94% (95%CI:0.89～0.97).The test of heterogeneity for specificity and specificity of CT for blunt injuries showed that there were statistical heterogeneity among studies included in the Meta-analysis(â… ~2=52.5%and 56.5%) and for penetrating the results were the same(â… ~2=61.5%and 79.9%),which indicated that random effect model was suitable for the synthesis.Conclusions:1.A high index of suspision should be kept on the basis of understanding the characteristics of the regularity of this disease,and the mechanism and condition of trauma should be evaluated carefully.In order to avoid missed diagnosis,frequent X examination preoperatively is necessary for increasing the rate of preoperative diagnosis in the absence of an immediate indication for surgery.Careful exploration should be done if emergency operation was done.2.X examination preoperatively and the blunt trauma type were two independent predictors for increasing the rate of preoperative diagnosis.The ablity of the performance of X-ray improves on the basis of the time allowing frequent examination and the intrathoracic visceral herniation imaging sign.3.Both sever associated injuries preoperatively and sever complications after-operatively are the direct reasons for mortality.7 factors identified by univariate analysis were different statistically for prognosis,which are ISS,volume of transfusion,shock symptoms when admitted to hospital,after-operative severe complications,age,sever associated injuries,and plenic injury scale.Adequate attention should be given to these patients associated with such indicators and do our best to eliminate or reduce the influence of the risk factors.4.Meta-analysis demonstrates that 3 factors can influence mortality significantly, which are shock symptoms when admitted to hospital,after-operative severe complications and sever associated injuries,and the remaining 5 factors are not risk for mortality statiscally,which are the type of trauma,sex,the side of ruptured diaphragm,the stage of disease progress,the time of diagnosis. Adequate attention should be given to these patients with the 3 risk factors, because they faced a great risk of death.5.Meta-analysis demonstrates that CT has a high diagnostic efficacy and it can be used as a reliable and stable non-invasive tool for diagnosing DI and TDH.CT reveals a good performance using all the different imaging sign,which are collar sign,diaphragmatic defect sign,intrathoracic visceral herniation sign,dependent viscera sign,and segmental diaphragmatic thickening sign.At practice,the comprehensive utilization of all the imaging sign should be fulfilled in order to improve the rate of preoperative diagnosis.There are no differences of CT performance between the diaphragm left-sided and right-sided injuries and the same for bunt and penetrating injuries.6.By merger and synthesis of the results of the same kind studies,meta-analysis can increase the sample size,enhance the statistic efficiency.In particular,when the results of different studies are not consensus or there are not statistic significance in results,meta-analysis could obtain statistic analytic result that approaches the real situation.7.The results of this studies would advantage the systematic evaluation of the patients with DI and TDH in the future clinic work,accurate judement of the diagnosis and prognosis for the patients,and the formation of theory for improving the outcome of treatments.Therefore,this study is helpful for the the clinic work in the future. |
