Quantification of “bacterial load” in late-onset neonatal sepsis using molecular techniques

Abstract

Premature infants are at increased risk of bacterial sepsis. Few studies of neonatal infection have quantified the degree of “bacterial load”. Traditionally, bacterial numbers have been estimated by measuring colony-forming-units per ml (CFU/ml), using conventional culture methods. Modern molecular techniques allow direct estimation of bacterial numbers in preterm infants with late-onset sepsis. Levels of bacterial 16S ribosomal RNA (rRNA) can potentially be used to quantify metabolically active bacteria in clinical samples. Objective To estimate the “bacterial load” in infants presenting with late-onset bacterial sepsis (>48 h old) using quantitative molecular methods. Methods A single-centred, observational study was conducted in all infants greater than 48 h old with clinically suspected sepsis. Blood samples (0.5 ml) were taken for a calibrated TaqMan reverse transcriptase polymerase chain reaction (RT-PCR) in addition to normal infection investigations. This assay allows quantitative detection of bacterial 16S rRNA and rDNA. Clinical and laboratory data were recorded at the time of sampling. Copy numbers (rRNA and rDNA) were calculated and corrected for background genetic material to estimate bacterial load. Results 100 samples were taken from 125 babies at a median age of 14 days. Median (interquartile) birth weight was 1.09 (0.81–1.45) kg and gestation was 27.5 (23–40) weeks. The rate of clinically diagnosed late-onset infection was 33%. The mean calculated and corrected copy number estimate (bacterial rRNA+rDNA) was 1.51×107/ml of blood (range 8.91×104 to 9.77×1010). Estimated copy number correlated positively with white blood cell count, neutrophil number and C-reactive protein and negatively with platelet count, lymphocyte count, pH and base deficit. There was no correlation with birth weight or gestational age. Conclusion Bacterial load in late-onset sepsis can be estimated by quantitative RT-PCR using a 16S rRNA assay. Direct comparison to estimates of CFU/ml was not possible because of the variability of rRNA numbers per bacterial cell. However, the potential of this target as a quantitative biomarker of bacterial infection was demonstrated. Increasing numbers of bacteria appear to stimulate greater immune response. Further study is necessary to place these values in context with other age groups and disease severity.