Introduction
Staphylococcus aureus toxins are implicated in sudden infant death syndrome (SIDS) as there is an inverse relationship between
incidence and circulating anti-toxin levels. A recent retrospective review of autopsy reports found that S. aureus was isolated more commonly
from cases of SIDS compared to an explained non-infection group. Another study detected the toxins in tissues of over 50% of cases from three
different countries. However, these toxins could be due to contamination at sampling. We have hypothesised that toxins not neutralised by
IgG will be eliminated to the urine and detected. Here we demonstrate a temperature dependant pattern of secretion which does not support
contamination as a source of toxins.

Methods
TSST-1 and SEB producing strains of S. aureus were cultured for 24h at 41°C, 37°C, 30°C, 22°C or 4°C in either Brain Heart Infusion
broth (BHI) or clean catch mid-stream urine. The OD600 of the suspensions were measured prior to centrifugation at 900g. Toxin levels in the
resulting media supernatants were determined by ELISA, gel electrophoresis and western blotting (WB).

Results
Maximum bacterial growth was observed at 30°C and 37°C while it was diminished at 22°C and 41°C and undetectable at 4°C. In contrast,
maximal TSST-1 release occurred at 41°C (0.10µg/ml) with less at 37°C (0.052µg/ml), while at 4°C, 22°C and 30°C TSST-1 was undetectable.
Similar results were obtained for SEB.

Conclusions
Contamination of urine samples stored at room or refrigerated temperature does not result in the release of biologically significant
levels of toxin. These results support the interpretation of our previous findings that the detection of S. aureus toxins in infant urine results from
a transient bacteraemia rather than contamination. ELISA and WB for bacterial toxins in body fluids taken at autopsy can be used to distinguish
between genuine infection with S. aureus and contamination.