Abstract: Objective To investigate the predictive value of serum proadrenomedullin(pro-ADM)level in acute cerebral infarction patients with stroke-associated pneumonia(SAP). Methods A total of 100 patients with acute cerebral infarction were selected and divided into SAP group(n=43)and non-SAP group(n=57)according to the clinical symptoms and microbiological culture results, and the time to diagnosis of SAP was recorded. The levels of serum pro-ADM were measured by ELISA on 1^st, 3^rd and 5^th day after admission. The predictive value of pro-ADM for SAP was assessed by receiver operating characteristic(ROC)curve analysis. The results of patients diagnosed as SAP on the 3^rd and 5^th day before the assay were not included in the predictive analysis. The pro-ADM levels of patients with NIHSS score≥10 at admission in SAP group and non-SAP group were compared. Multivariable logistic regression was applied to investigate risk factors for the progression of SAP. Results On the 1^st, 3^rd and 5^th day after hospital admission, the pro-ADM levels were(1.74±0.39),(1.44±0.44)and(1.56±0.47)nmol/L respectively in the SAP group, which were significantly - higher than(1.30±0.33),(1.23±0.42)and(1.13±0.39)nmol/L in non-SAP group(P<0.05 or P<0.01). For diagnosis of SAP, the area under the ROC curve( AUC)for pro-ADM on 1^st, 3^rd and 5^th were 0.807(95% CI was 0.724 to 0.890), 0.636(95% CI was 0.521 to 0.751)and 0.760(95% CI was 0.645 to 0.875), respectively. On the 1^st and 5^th day, the pro-ADM levels of patients with NIHSS score≥10 at admission were(1.71±0.37)and(1.55±0.49)nmol/L in SAP group, which were significantly higher than(1.48±0.34)and(1.24±0.34)nmol/L in non-SAP Group(P<0.05). The multivariable logistic regression analysis showed that a higher NIHSS score and a higher serum pro-ADM level on 1^st day after hospital admission were the risk factors for SAP. Conclusion The detection of serum pro-ADM level in the early stage of acute cerebral infarction may be helpful to predict the occurrence of SAP.