• Sergio Serrano-Gomez, Gabriel Burgos-Angulo, Daniela Camila Niño-Vargas, María Eugenia Niño, María Eugenia Cárdenas, Estephania Chacón-Valenzuela, Diana Margarita McCosham, Juan Sebastián Peinado-Acevedo, M Marcos Lopez, Fernando Cunha, Antonio Pazin-Filho, Ramses Ilarraza, Richard Schulz, and Diego Torres-Dueñas.
• Department of Public Health, Medicine Program, Hospital Universitario de Santander, Bucaramanga, Santander, Colombia.
• J Intensive Care Med. 2020 Jan 1; 35 (1): 95-103.

PurposeOver 170 biomarkers are being investigated regarding their prognostic and diagnostic accuracy in sepsis in order to find new tools to reduce morbidity and mortality. Matrix metalloproteinases (MMPs) and their inhibitors have been recently studied as promising new prognostic biomarkers in patients with sepsis. This study is aimed at determining the utility of several cutoff points of these biomarkers to predict mortality in patients with sepsis.Materials And MethodsA multicenter, prospective, analytic cohort study was performed in the metropolitan area of Bucaramanga, Colombia. A total of 289 patients with sepsis and septic shock were included. MMP-9, MMP-2, tissue inhibitor of metalloproteinase 1 (TIMP-1), TIMP-2, TIMP-1/MMP-9 ratio, and TIMP-2/MMP-2 ratio were determined in blood samples. Value ranges were correlated with mortality to estimate sensitivity, specificity, positive predictive value, negative predictive value, and area under the receiving operating characteristic curve.ResultsSensitivity ranged from 33.3% (MMP-9/TIMP-1 ratio) to 60.6% (TIMP-1) and specificity varied from 38.8% (MMP-2/TIMP-2 ratio) to 58.5% (TIMP-1). As for predictive values, positive predictive value range was from 17.5% (MMP-9/TIMP-1 ratio) to 70.4% (MMP-2/TIMP-2 ratio), whereas negative predictive values were between 23.2% (MMP-2/TIMP-2 ratio) and 80.9% (TIMP-1). Finally, area under the curve scores ranged from 0.31 (MMP-9/TIMP-1 ratio) to 0.623 (TIMP-1).ConclusionAlthough TIMP-1 showed higher sensitivity, specificity, and negative predictive value, with a representative population sample, we conclude that none of the evaluated biomarkers had significant predictive value for mortality.

31 keywords...

hide…