Performance of 16S rRNA/18S rRNA PCR and amplicon sequencing compared to conventional culture methods for the identification of bacterial/fungal pathogens in diagnostically challenging infections

Astrid Muyldermans*1, Kurt Beuselinck1, Katrien Lagrou1, Stefanie Desmet1

Author address: 

1 Department of Laboratory Medicine; Department of Microbiology and Immunology, University Hospitals Leuven; KU Leuven - University of Leuven, Leuven, Belgium


Background: Identification of causative micro-organisms (MO) is important for adequate management and directed therapy of infections. 16S and 18S rRNA PCR with amplicon sequencing are broad-spectrum methods detecting bacteria and fungi. In this retrospective study we evaluated the performance of these PCRs compared to conventional culture methods.

Materials/methods: Samples received between October 2013 and October 2018 for both culture and 16S rRNA (excluding samples from primary non-sterile sites) and/or 18S rRNA PCR were included. DNA extraction, PCR and sequence analysis were performed with UMD-Universal kit (Molzym).

Results: 16S rRNA PCR was performed on 582 samples [heart valve (n=212); other biopsies (n=182); CSF (n=51); joint aspirates (n=43); other body fluids (n=94)], 18S rRNA PCR on 140 samples [biopsies (n=63); BAL (n=37); other body fluids (n=40)].


PCR and culture were concordant in 305/582 (52.4%) samples. Additional MOs, not detected by culture, were found in 226/582 (38.8%) samples (Figure 1). In 51/582 (8.8%) samples, MOs identified by culture were not detected by PCR, including Candida spp. (n=4) and Aspergillus spp. (n=2) not targeted by 16S rRNA PCR.

In 127/212 (59.9%) heart valves, additional MOs were detected. In only 8/212 (3.8%) valves, PCR missed MOs.

In 46/182 (25.3%) non-heart valve biopsies and 53/188 (28.2%) body fluids, PCR detected additional MOs. MOs were missed in 41/182 (22.5%) non-heart valve biopsies (mainly coagulase-negative staphylococci which are potentially contaminants) and 2/188 (1.1%) body fluids.


18S rRNA PCR was positive in 29/140 samples. Compared to culture, Aspergillus species were additionally detected by PCR in 5/140 (3.6%) samples. In 3/140 samples, PCR was positive but identification was not possible. In one sample, a potential pathogen was missed (Rhizomucor pusillus in BAL).

Conclusions: 16S rRNA PCR provides an increased detection rate of bacteria, with highest number of additional detections in heart valves. However, conventional culture should still be performed in parallel for optimal sensitivity and susceptibility testing.

Only little gain in detection rate of fungi by 18S rRNA PCR is observed, nevertheless it can be crucial in identifying pathogens in culture-negative infections.

Figure 1. 16S rRNA PCR with amplicon sequencing compared to bacterial culture.




Image icon image.png719.26 KB

abstract No: 


Full conference title: 

European Congress of Clinical Microbiology & Infectious Diseases 2019
    • ECCMID 29th (2019)