The outbreak was triggered after a 43-year-old woman was admitted to the intensive care unit at the National Institutes of Health (NIH) research hospital in Maryland, USA. She had been diagnosed with an infection of the bacterium Klebsiella pneumoniae and was immediately put into quarantine.
K. pneumoniae kills around half of all people diagnosed with it, and resists most antibiotics. It can also be spread by carriers who show no symptoms of infection.
The hospital used a technique known as 'enhanced contact isolation', enforcing the use of masks and gloves at all times, disinfecting all equipment and disposing of any medical supplies which could not be thoroughly cleaned.
However, three weeks after the woman was discharged, a new case of K. pneumoniae was diagnosed in a 34-year-old cancer patient who had never had any contact with her.
Doctors at the hospital were unsure whether the two cases were related. For other bacteria, the hospital would use a technique called pulsed-field gel electrophoresis to differentiate between strains. But due to the nature of K. pneumoniae, that technique was unhelpful.
Despite the hospital's best efforts, the infection spread to another 15 patients over the course of five months. Initially, it was confined to the intensive care unit, but it soon spread to the rest of the hospital. By now, the hospital was 'cohorting', ensuring staff treating the infected patients were not working elsewhere, and employing extreme disinfecting techniques, even fumigating entire rooms with hydrogen peroxide gas.
Dr Julie Segre, a genome researcher at NIH, suggested that sequencing the entire genome of the first patient's bacteria would allow a comparison with the genome sequences of bacteria from other patients, mapping the spread of the disease.
Sequencing demonstrated that the first patient was responsible for all subsequent infections, transmitting the bacteria from her lung and throat on three separate occasions.
'It is very difficult to contain an epidemic once these kinds of organism are introduced into the hospital environment', said David Henderson, NIH Clinical Centre's deputy director for clinical care. 'They become endemic and part of the hospital flora'.
The bacteria from the patient's lungs had a difference of seven base-pairs out of six million from those her throat, allowing the researchers to make a detailed analysis how it had spread.
Every patient in the hospital was also tested for the bacteria, revealing eight carriers who showed no symptoms, explaining the apparent gaps between infected patients. The contamination was eventually tracked to the plumbing connected to the first patient's isolation unit, and a ventilation machine which had not been successfully disinfected.
'We had never done this type of research in real time', Dr Segre said.
The sequencing pointed to a source of infection that was unforeseen, and showed that the bacteria could survive undetected for an extremely long time. The research could lead to future use of genome sequencing in hospitals to prevent the spread of other infectious diseases. In the USA, more than 99,000 people die every year from hospital-acquired infections.
'It's great to see this study', Professor Sharon Peacock, from the University of Cambridge, UK, told New Scientist. 'It's going to lead to a step change'.
Currently, most hospitals in the UK do not have the expertise or resources to carry out the analysis in-house and some researchers suggest an online database of bacterial genomes would help in the future.
'The quality of answer will only be as good as the database', Professor Peacock commented.