THE wait must be agonising. As H5N1 bird flu moves across Asia and into Europe, dozens of families, from China to Turkey, are spending days waiting for a verdict from government laboratories to confirm whether or not their loved one has the virus. And even then the diagnosis has sometimes proved to be wrong.
We can do better. There are several technologies in the pipeline that could vastly speed up the diagnosis of respiratory ailments, and detect the precise strain causing a flu infection in minutes. Yet national pandemic plans have so far neglected improvements in diagnosis, and the public health establishment appears reluctant to move away from the reliable but plodding lab-based tests now in use.
If a pandemic were to strike, health authorities would be faced with large numbers of people clamouring for stockpiled antiviral drugs such as Tamiflu. Those limited supplies will have to be given only to people who might benefit from them, and not to those with the bacterial pneumonia that often follows flu, or the 鈥渨orried well鈥�, says Graeme Laver of the Australian National University in Canberra, who helped discover the antiviral class that includes Tamiflu. And to be effective, the drugs must be taken very early in the course of the illness. 鈥淭here should be dispensaries where people get a fast diagnostic test before they get the drug,鈥� he says. So far, no country even has such a facility on the drawing board.
Advertisement
There are fast dipstick tests, similar to pregnancy tests, which can detect whether someone has flu, but these cannot identify which strain. That may not matter when a pandemic is at its height, but before it strikes it鈥檚 important to know exactly what virus is causing any outbreak of a flu-like illness. At present, that takes time. The tests carried out in approved government labs convert the flu virus鈥檚 RNA genome to DNA, and amplify the DNA using the polymerase chain reaction (PCR). The amplified DNA is then incubated with genetic sequences from various flu strains to see which bind to it. 鈥淭hat can all be done within 24 hours,鈥� says flu specialist Albert Osterhaus of Erasmus University in Rotterdam, the Netherlands.
But transporting samples to a specialised lab takes time, especially if the lab is in another country. And bureaucracy can add to the delays. Last October the first suspected H5N1 samples from birds in Romania were delayed by a day on the way to a British lab while the paperwork required to meet European Union regulations on transporting hazardous material was prepared.
Yet speed is essential. If the H5N1 strain mutates and starts to be transmitted from person to person, treating all those who have been exposed to it with antiviral drugs might stop it spreading beyond the first cluster. If diagnosis takes days, infected people might leave the area before results are available, taking the virus with them.
Last month, flu experts met in Tokyo to discuss how to make such containment work. 鈥淭here was a real consensus that diagnostic labs need higher standards, training and equipment,鈥� says Keiji Fukuda of the World 午夜福利1000集合 Organization鈥檚 global influenza programme. But those measures may be of limited use if there is no quick way to get samples from the initial cluster to the lab.
The technology developers say they will be able to avoid delays entirely with diagnostic tools that can be used at the scene of an outbreak. 鈥淲e intend for our kits to be used by paramedics,鈥� says Stefano Lopriore of STMicroelectronics in Geneva, Switzerland, which last month announced that by next year it will produce a 鈥渓ab-on-a-chip鈥� able to diagnose a flu strain as accurately as standard lab techniques, but in a fraction of the time.
The lab-on-a-chip consists of an array of DNA sequences from different known strains. A sample is taken from a patient and the RNA of any flu virus in it is converted to DNA and amplified using PCR in the same way as standard tests. But before the sample is tested, the DNA is attached to a fluorescent molecule, so when it binds to the 鈥渃apture鈥� DNA on the chip it can be spotted by an automatic reader. This pinpoints where the glow is coming from and so identifies the strain.
Originally developed to chart changes in gene expression over time, such 鈥渕icro-array鈥� devices are increasingly being used to identify pathogens. A device of this type helped identify the SARS virus in 2003. One limiting factor has been the binding time. Kathy Rowlen and colleagues at the University of Colorado, Boulder, have developed a micro-array that can identify a strain of flu with a success rate comparable to or better than standard tests, according to team member Michael Townsend. But the device still takes 11 hours to produce a result, as the viral RNA still has to be turned into DNA and amplified by PCR, then incubated on the chip. The whole thing is not much quicker than standard lab tests.
Townsend says the time might be cut dramatically if it was possible to put the viral RNA directly onto the chip, cutting out the DNA and amplification steps. One problem is that this would make the samples much smaller, so the team is investigating the use of ultra-sensitive signal detection in order to pick up any faint fluorescence from the captured flu genes. They hope to make an array costing around $30.
鈥淭here are several technologies that could detect the precise flu strain in minutes鈥�
This contrasts significantly with a micro-array for diagnosing flu launched by Combimatrix, based near Seattle, Washington, in December. Although capable of identifying a flu strain in 4 hours, it costs around $700.
STMicroelectronics, meanwhile, plans to slash the diagnosis time of its device to 1 hour by carrying out the PCR stage on the chip itself. The sample is amplified by a micro-PCR chamber and then immediately exposed to the capture-DNA sequences, where it binds with any matching strain.
The cost of lab-on-a-chip devices could prove prohibitive, particularly for developing countries carrying out extensive testing programmes, as each one must be disposed of after a few tests. A reusable diagnostic device would be a good alternative. This is the approach being pursued by Isis Pharmaceuticals of Carlsbad, California, which is using technology originally developed by the Pentagon鈥檚 Defense Advanced Research Projects Agency for spotting bioterrorist attacks (New Scientist, 19 September 1998, p 42).
TIGER, or Triangulation Identification for the Genetic Evaluation of Risks, first amplifies the unknown virus with PCR in the same way as the lab-on-a-chip devices. It then squirts a tiny amount into a mass spectrometer, a device used to separate molecular fragments according to their mass and charge. This picks up subtle differences in the DNA, generating a fingerprint for each pathogen within 30 seconds. It has already been used to track an outbreak of a bacterial disease on a naval base, and has been bought by the forensics lab at the US Department of Homeland Security.
鈥淚n the next two weeks we will perform a huge validation trial for flu,鈥� says David Ecker of Isis, who helped develop the TIGER system. The machine costs $350,000, but will pay for itself after just a few thousand tests, he says. TIGER also reports every bug in the sample, so if you don鈥檛 have flu it can tell you what you do have: bacterial pneumonia, for example. The portable device can be used wherever it is needed, and Ecker says it is at least as accurate as laboratory PCR tests.
But convincing public health officials to abandon their tried and tested 鈥� if painfully slow 鈥� approach may be no easy task. Fukuda is wary of the idea of non-specialists carrying out diagnosis outside government labs, especially in the early days of a looming pandemic. Exploiting new technology to shave two days off the test would be nice, he says. 鈥淏ut it won鈥檛 help if we can鈥檛 be sure of the result.鈥�
