Rapid detection and typing of live bacteria from human joint fluid samples by utilizing an integrated microfluidic system

Abstract

Periprosthetic joint infection (PJI) is one of the most dreading complications that hinder the merits of an arthroplasty. A prerequisite for treatment of the above procedure is rapid detection of live bacteria to prevent its recurrence and proper choice of antibiotics. Conventional culture methods are time-consuming and associated with a high false negative rate. Amplification of bacterial genetic materials requires a tedious process but is associated with a high false positive rate. An integrated microfluidic system capable of molecular diagnosis for detecting live bacteria was reported in our previous work. However, the system could not provide detailed information about infectious bacteria for the subsequent antibiotic choices. Furthermore, it took at least 55min to finish the entire process. In this work, a microfluidic platform using ethidium monoazide (EMA) which can only penetrate into dead bacteria is presented for live bacteria detection and typing within a short period of time (30min for the detection of live bacteria and another 40min for the typing of bacteria strains). We tested the proposed system by using human joint fluid samples and found its limit of detection for bacterial detection equal to 102CFU (colony formation unit) for live bacteria detection with gold nanoparticle probes and 102–104CFU for typing bacteria by an on-chip polymerase chain reaction. The whole procedure of the integrated microfluidic system is automated with little human intervention. Moreover, this is the first time that sequential live bacteria detection and typing are demonstrated on the same microfluidic platform. Based on the promising results, the proposed system may become in the near future an auxiliary tool for immediate medical decision and choice of antibiotics in routine arthroplasties or PJI's.