Small-molecule correlates of infection precede infection diagnosis in breast implant reconstruction patients

BACKGROUNDInfection is an important complication of implanted devices and prosthetics. Identifying infections sufficiently early to salvage implants and avoid reconstructive failure is a persistent medical challenge.METHODSTwo cohorts of women 21 years and older undergoing breast implant reconstruction were recruited. Seroma fluid (82 breasts, 70 patients) was collected upon implant removal for infectious or noninfectious causes. Postimplantation drain fluid (100 samples, 44 breasts, 32 patients) was collected at routine visits prior to implant removal. A liquid chromatography/mass spectrometry-based metabolomic approach was used to identify infection correlates.RESULTSIn seroma fluid specimens, infection was associated with a diverse set of small molecules, including acetylated polyamines, defensins, glucosyl-sphingosine, and several peptide-like features (all P < 0.001, diagnostic areas under the receiver operating curve 0.82-0.93). Notably, a subset of these markers were significantly elevated (P < 0.05) in postimplantation drain fluid before recorded infection symptoms and diagnosis. Pseudomonas aeruginosa and its specialized exometabolites in drain specimens were also associated with subsequent P. aeruginosa infections.CONCLUSIONTissue fluid from infected patients has a distinctive metabolome reflecting human and bacterial physiologic processes that often precede clinical diagnoses. A diagnostic based on these findings has potential to improve patient outcomes through early recognition of infection.FUNDINGThis work was supported by U54CK000609 from the CDC and by an unencumbered research gift from Sientra. Metabolomic approaches were supported by NIH grants R01 DK125860 and R01 DK111930.