| BackgroundSerum albumin probably was a neuroprotective agent. Except for a large sampleprospective studie on the association between serum albumin and the outcome of acuteischemic stroke, others were all small size with short time of follow up. The meansample size of these studies was merely350. The evidence was not sufficient,although, these studies provided a clue of the association. Serum globulin may affectthe progress of acute ischemic stroke, which took part in infection and inflammation.Studies on the association between serum albumin and the outcome of acute ischemicstroke was rare, the sample size from17to329, with an average of146. However, theassociation between admission serum albumin, serum globulin and the prognosis ofacute ischemic stroke is inconclusive. Therefore a large sample, multi-center cohortstudy to provide scientific evidence for the association of albumin, globulin andprognosis of acute ischemic stroke was required.Purpose1. To explore the effects of admission serum albumin, globulin on dischargedoutcomes among patients with acute ischemic stroke.2. To explore the interaction of hypoalbuminemia and hyperglobulinemia withhypertension, hyperglycemia and hyperlipidemia.3. To explore the predictive effect of admission serum albumin, globulin andconventional factors on discharged outcomes among patients with Acute IschemicStroke.4. To provide clinical epidemiology evidence for the secondary and thirdprevention. MethodsSubject: In a non-concurrent cohort study from June1st2009to May31,2013, wecontinuously included12185patients with acute ischemic stroke from Fuxin CentralHospital of Liaoning Province, Xinganmeng People’s Hospital of Inner MongoliaAutonomous Region, Affiliated Zhongshan Hospital of Dalian University and AffiliatedHospital of Chengde Medical College.Methods: Trained and qualified investigators collected data on the information ofdemographic characteristics, lifestyle, admission serum albumin, serum globulin andother clinical laboratory tests. Outcome was measured with the modified Rankin scale(mRS) when subjects were discharged. Good outcome was defined as0≤mRS≤2, andpoor outcome was defined as3≤mRS≤6.Statistical analysis: EpiData3.1software was used to establish the database, and allquestionnaires were input after checked by two individuals. SAS9.2software was usedfor statistical analysis. The t-test, chi-square and Wilcoxon-log rank test were performedto test the diffrence of general information between the good and the poor outcome.Binary Logistic regression analysis and ordinal Logistic regression analysis wereconducted to estimate the association between admission serum albumin, serumglobulin and the discharged outcome of acute ischemic stroke and and calculated therelative effects of various combinations by exposure factors of hypoalbuminemia,hyperglobulinemia, hypertension, hyperglycemia and hyperlipidemia. The additiveinteraction was evaluated by the interaction calculation sheet which made by Andersson,et al. Receiver Operating Characteristic Curve was used to estimate the predictive effectof admission serum albumin, globulin and conventional factors on discharged outcomesamong patients with Acute Ischemic Stroke. All tests were two-sided test and asignificance level of0.05.Results1.12185patients with acute ischemic stroke were included in the analysis fromfour clinical centers. Prevalence of disability and case fatality rate were17.60%and1.94%, respectively. Of12185patients with acute ischemic stroke,7670(62.95%)were male,4515(37.05%) female and the ratio of male to female was1.70:1. Allthese subjects were aged from40to99with62.27±11.38years old. The rate ofhypoalbuminemia and hyperglobulinemia were12.97%and29.39%, respectively.2. Compaired with the patients with good discharge outcome, subject with poor discharge outcome tended to be older, with shorter time from onset to admission, loogerstay in the hospital, with higher history rates of of diabetes, heart disease, atrialfibrillation, recurrence, higher levels of SBP, DBP, fasting plasma glucose, higher rateof hypertension, hyperglycemia, High TG, Low HDL-C and hyperlipidemia (all P<0.05).Patients from different hospital had different rate of poor discharge outcome, andsubjects with poor discharge outcome had lower level of serum albumin, higher level ofserum globulin and higher rates of hypoalbuminemia and hyperglobulinemia (allP<0.05).3. Serum albumin was divided into four groups according to the quartile. Therates of poor outcome in these four groups were26.98%、19.18%、16.09%and16.03%,respectively. The rates of poor outcome with serum albumin in the second, third andthe forth quartile were lower than that in the first quartile (P<0.05). Serum globulinwas divided into four groups according to the quartile. The rates of poor outcome inthese four groups were15.50%,16.92%,19.40%and26.24%, respectively. The ratesof poor outcome with serum albumin in third and the forth quartile were higher than thatin the first quartile (P<0.05).4. Compaired with the first quartile of serum albumin, the proportion of smallermRS with serum albumin in the second, third and the forth quartile were higher thanthat in the first quartile (P<0.05) and with the quartile of serum albumin elevated, therewas a lift shift of patients on the axis of mRS. Compaired with the first quartile ofserum globulin, the proportion of smaller mRS with serum globulin in the second, thirdand the forth quartile were lower than that in the first quartile (P<0.05) and with thequartile of serum globulin elevated, there was a right shift of patients on the axis ofmRS.5. In the Binary Logistic regression analysis, after adjusting comfouding factors,compaired with the first quartile of serum albumin, the OR and95%CI in the second,third and the forth quartile were0.712(0.627-0.810),0.604(0.527-0.692) and0.619(0.539-0.712), respectively. With1g/L serum albumin increased, the risk of pooroutcome increased4.6%(OR:0.954，95%CI:0.944-0.964). The risk of poor outcomeamong patient with hypoalbuminemia was1.865times as high as patient withouthypoalbuminemia, OR and95%CI was1.865(1.642-2.118). Compaired with the firstquartile of serum globulin, the OR and95%CI in the third and the forth quartile were 1.325(1.149-1.528) and1.721(1.497-1.979), respectively. With1g/L serum globulinincreased, the risk of poor outcome increased2.7%(OR=1.027，95%CI:1.018-1.037).The risk of poor outcome among patient with hyperglobulinemia was1.488times ashigh as patient without hypoalbuminemia, OR and95%CI was1.488(1.343-1.648).6. In the ordinal Logistic regression analysis, after adjusting comfouding factors,compaired with the first quartile of serum albumin, the OR and95%CI in the second,third and the forth quartile were0.748(0.682-0.821),0.696(0.633-0.765) and0.646(0.586-0.712), respectively. With1g/L serum albumin increased, the risk of atleast1score elevated in mRS was0.962(0.955-0.968). The risk of at least1scoreelevated in mRS among patient with hypoalbuminemia was1.679(1.522-1.852)compaired with patients in normal level of serum albumin. Compaired with the firstquartile of serum globulin, the OR and95%CI in the second, the third and the forthquartile were1.118(1.018-1.227),1.185(1.078-1.303) and1.391(1.263-1.532),respectively. With1g/L serum globulin increased, the risk of at least1score elevatedin mRS was1.019(1.012-1.026). The risk of at least1score elevated in mRS amongpatient with hyperglobulinemia was1.247(1.158-1.342), compaired with patients innormal level of serum globulin.7. There were significantly interaction between hypertension andhypoalbuminemia on discharged poor outcome among acute ischemic stroke patients inadditive model with RERIs of0.546, the APs of20.3%and the S index of1.477. Therewere significantly interactions between hyperglycemia, hypoalbuminemia andhyperglobulinemia on discharged poor outcome among acute ischemic stroke patients inadditive model with the RERIs of0.894and0.910, the APs of27.1%and34.6%and Sindexs of1.637and2.261, respectively.8. The areas under receiver operating characteristic curve (area under the curve,AUC) for the second model containing serum albumin and conventional factors(AUC=0.6960), the third model containing serum globulin and conventional factors(AUC=0.6910), the forth model containing serum albumin, globulin and conventionalfactors (AUC=0.7014), the fifth model containing the ratio of serum albumin toglobulin and conventional factors (AUC=0.6985), were significantly (all P<0.05) largerthan the first model containing only conventional factors (AUC=0.6855). As expected,the AUC for the forth model containing serum albumin, globulin and conventional factors was significantly (P=0.0128) larger than the fifth model containing the ratio ofserum albumin to globulin and conventional factors.9. In the ordinal Logistic regression analysis, after adjusting admission NIHSSscore and other comfouding factors, compaired with the first quartile of serum albumin,the OR and95%CI in the second, third and the forth quartile were0.839(0.707-0.997),0.736(0.617-0.878) and0.787(0.648-0.956), which respectively remained significant.With1g/L serum albumin increased, the risk of at least elevated1score in mRS were0.959(0.944-0.974). The risk of at least elevated1score in mRS among patient withhypoalbuminemia was1.652(1.368-1.996). Compaired with the first quartile of serumglobulin, the OR and95%CI in the secomd, the third and the forth quartile were1.313(1.111-1.552),1.339(1.120-1.601) and1.851(1.546-2.216), respectively. With1g/Lserum globulin increased, the risk of at least elevated1score in mRS were1.016(1.006-1.027). The risk of at least elevated1score in mRS among patient withhyperglobulinemia was1.523(1.313-1.768).Conclusion:1. Elevated serum albumin at admission decreased the risk of discharged pooroutcome among patients with acute ischemic stroke.2. Elevated serum globulin at admission increased the risk of discharged pooroutcome among patients with acute ischemic stroke.3. Interaction between hypertension and hypoalbuminemia, hyperglycemia andhypoalbuminemia, hyperglycemia and hyperglobulinemia probably increase the risk ofdischarged poor outcome.4. The predictive effect of admission serum albumin, globulin and conventionalfactors on discharged outcomes among patients with acute ischemic stroke was stronerthan the ratio of serum albumin to globulin and conventional factors. |
PDF Full Text Request