[Application of proteomic models for the diagnosis of inflammatory and non-inflammatory preterm delivery].

Objective: To establish the serum proteomic models for the identification of premature delivery with inflammation or non-inflammation.

Methods: The laboring patients from 2008 to 2010 at our hospital were divided into 5 groups according to placental pathology, including inflammatory preterm group, non-inflammatory preterm group and blind test group (n = 50 each). The control group was normal full-term. The preterm models with or without inflammation were established by the methods of SELDI-TOF-MS (matrix-assisted laser desorption/ionization-time of flight-mass spectrometry) and bioinformatics. And statistical analysis was performed after a blind test. Then differential protein fingerprints were compared and analyzed.

Results: A total of 36 different proteins were harvested after a comparison of inflammatory preterm and control groups. The model was established by eight markers. The rates of specificity, sensitivity, positive predictive value and negative predictive value were 80.85%, 84.44%, 93.33% and 46.67% respectively. Fifteen different proteins existed between non-inflammatory preterm and control groups. There were 4 marker proteins. The rates of specificity, sensitivity, positive predictive value and negative predictive value were 75.68%, 64.44%, 80.00% and 66.67% respectively. Different marker proteins existed between the identification models of inflammatory preterm labor and inflammatory full-term.

Conclusion: Different serum proteomic models may be used for the diagnosis of preterm labor and the differentiation of preterm labor with or without inflammation. And different proteins are expressed during different stages of pregnancy with chorioamnionitis.