Is bacteria cause of cot death?

“Cot death linked to common bacteria” is the headline in the Daily Mirror today. Researchers have pinpointed two common bacteria that may contribute to cot deaths. The Mirror and other newspapers report that postmortems carried out in 500 babies who died unexpectedly found high levels of Staphylococcus aureus and Escherichia coli .

The research behind the stories is a review of autopsy records for babies who died suddenly and unexpectedly. Researchers compared the results of microbiological tests between the babies whose cause of death was registered as unexplained, due to infection or due to non-infective causes. As the newspapers report, the study does not prove a link between these bacteria and cot death, but it suggests there may be some association. Importantly though, the results do not change the current recommendations to help prevent cot death; to avoid smoking around the baby, to put the baby to sleep on its back and to keep the baby at a comfortable temperature without covering its head.

Where did the story come from?

Dr Martin Weber and colleagues from Great Ormond Street Hospital and the Institute of Child Health at University College London carried out this research. The study was funded by the Foundation for the Study of Infant Deaths. It was published in the peer-reviewed medical journal: The Lancet .

What kind of scientific study was this?

The study was a retrospective case series, i.e. a review of autopsy records for 546 babies who died and had an autopsy to investigate sudden unexpected death in infancy (SUDI) at one specialist centre (Great Ormond Street Hospital) between 1996 and 2005. The researchers extracted data from the autopsy records relating to all organ systems (microscopic and macroscopic findings). The autopsies involved taking samples from all the organs and studying them under the microscope, as well as taking samples of fluids from various places to assess whether there were any bacteria, viruses or fungi present – i.e. microbiological samples. The final cause of death was classified from autopsy results as due to:

• infection (i.e. evidence of acute inflammation of cells normally associated with bacterial infection and deemed sufficient to be responsible for death);
• a non-infectious cause (e.g. head injury, heart disease or other causes);
• unknown cause (unexplained death after autopsy, excluding the results of microbial analysis).

The researchers excluded from the analysis autopsies with evidence of viral or non-bacterial infection. They also excluded autopsies in which cause of death was a non-infective one, but in which there was a secondary infection. From the original sample, 39 were excluded because they had a viral or secondary infection. Out of the remaining 507 for whom they were able to assess a “cause of death” from autopsy, 470 also had microbiological samples that they could use. These 470 autopsies were included in their analyses. Some of the babies who died from non-infectious causes, where the cause of death could be determined without microbiology, did not have samples taken, and so were not included in the analysis.

The researchers categorised the microbiological results into three categories: non-pathogens (i.e. bacteria that don’t usually cause illness), group 1 pathogens (i.e. disease-causing bacteria that normally have a defined focus of infection), and group 2 pathogens (i.e. disease-causing bacteria that can cause septicaemia without a focus of infection). They then compared the results of the microbiological findings between the babies with the different causes of death to see whether there were any differences between them.

What were the results of the study?

For the majority of the deaths the cause of death was “unexplained” (75%); 14% of deaths were due to non-infective causes and 11% were due to infection. Of the bacteriological samples taken, 73% were positive for microorganisms with the remaining 27% being “sterile”, i.e. no bacteria present.

When the researchers looked at the non-disease causing pathogens and the group 1 pathogens, there was no difference in the concentration of these found in babies who died of non-infective causes and those whose death was unexplained. However, there were more group 2 pathogens found in the unexplained death group compared with those in the non-infective cause group. Not unexpectedly, the babies who died from infection had the highest levels of group 2 pathogens. The greatest proportion of samples in babies with unexplained cause of death contained Staphylococcus aureus and Escherichia coli .

What interpretations did the researchers draw from these results?

The researchers conclude that their study suggests that microbes could be linked to some proportion of sudden, unexpected deaths in infants. They say that the reasons for this are unclear and should be investigated. Importantly, they say that the mere detection of disease-causing organisms does not prove they are the cause of death; in their own study they found group 2 bacteria present in a quarter of babies who had died from non-infective causes and, overall, about three quarters of samples were positive for some sort of bacterium.

What does the NHS Knowledge Service make of this study?

• By virtue of its design, this study cannot prove that the bacteria caused the death of the children in the “unexplained” group. The researchers themselves acknowledge that a problem with these findings will be their application and that it will be important to work out ways to distinguish between infections that actually caused death and those that are as a result of contamination or simple colonisation. They say that many of the isolates in their study could have been a result of colonisation during resuscitation for example, but that this doesn’t explain why babies with unexplained death should have more disease-causing bacteria than those who died from non-infective causes.

• It is possible that more microbiological samples were taken from babies with unexplained deaths in an attempt to find a cause, this would tend to increase the relative bacteriological yield in this group compared with babies who had died from clear bacterial or non-infectious causes.

• How these findings are put to use remains to be seen. Babies with unexplained death didn’t have the classic signs of infection, i.e. no cell inflammation etc. suggesting that if the pathogens were responsible for the babies’ deaths, it was by unknown mechanisms. It is known that these bacteria release toxins, however, they usually lead to classic infection symptoms such as inflammation and this was not seen in this study. 
• For now, this study is not conclusive and it does not prove that these bacteria cause sudden, unexplained death in infants. The researchers say that the presence of the bacteria may be an indicator of another cause of death (rather than be the cause of it themselves). For example, they suggest that “overheating or impaired infant arousal” which are already known to be associated with sudden unexplained death in infants.

The findings of this study do not alter the current recommendations to help prevent sudden infant death syndrome, i.e. to avoid smoking around the baby, to put the baby to sleep on its back, and to keep the baby at a comfortable temperature with its head uncovered.