The world last week seemed to edge closer to the brink of a flu pandemic. On 30 June, officials at the World Health Organization (WHO) revealed that they recently considered raising the threat level of a global pandemic, from the current 3 on a six-point scale, to 4 or even 5.

The scare was triggered a few weeks ago when several research groups visiting Vietnam filed preliminary reports that many people with mild cases of influenza — and those in contact with them — were testing positive for the deadly avian flu strain H5N1. This suggested that there was widespread human-to-human transmission of the virus.

Subsequent tests have so far failed to confirm this, and WHO spokesman Dick Thompson is keen to play down the incident. “It was just unpublished information provided to us in preliminary form that spurred an investigation,” he says. “We thought about upgrading the alert. We looked at it fast and strongly, and based on that decided not to upgrade.”

But take a closer look, and the picture in Vietnam is one of confusion rather than reassurance. The first signs of trouble came in May, with reports of small clusters of human cases of H5N1, including a rise in the infection of older people and an increase in milder cases — all signs consistent with the possibility that the virus had mutated to achieve improved, although still inefficient, human-to-human spread (see Nature 435, 390; 200510.1038/435390a).

Concern mounted in subsequent weeks as several international groups investigated the clusters using different methods, including the polymerase chain reaction (PCR), which amplifies DNA sequences, and western blots, which use antibodies to detect proteins. Despite using different tests, each of the teams reported that “substantial proportions” of the hundreds of people it had tested seemed to be infected with H5N1.

That led the WHO to consider upgrading the pandemic threat level to 4 (small, localized clusters of human infection) or 5 (large clusters of infection) — just one step away from a full-blown global pandemic. But first it asked an international team of experts, including Masato Tashiro, a virologist at the National Institute of Infectious Diseases in Tokyo, to retest many of the samples and some new ones, using the WHO's own PCR tests.

They found no evidence of clusters of human-to-human transmission. “This is good news,” Tashiro says, relieved that his worst suspicions weren't confirmed. But it remains unclear why the various groups got different results.

Close to danger? A young boy feeds ducks, a source of the H5N1 flu virus, in Vietnam. Credit: HOANG DINH NAM/AFP/GETTY IMAGES

Samples have now been sent to a WHO laboratory in Hong Kong for the last word in confirmation: antibody neutralization assays. These take time as they involve growing large amounts of the virus for analysis, but a firm conclusion is expected by the end of the month.

In the meantime, the WHO is holding off raising the alarm. “Because of the consequences of such a change, the WHO is following a cautious approach,” it said in a statement last week. Pushing the level to 4 for the first time would mean deploying the international stockpile of antiviral drugs to try to contain or stamp out the spread, and would probably result in countries restricting travel to Vietnam.

But Tashiro remains concerned that he and his colleagues didn't have enough time to check all of the clinical and epidemiological information associated with the initial lab samples. Follow-up work is also complicated by the fact that recovered patients have now returned home, making it hard to track down people they might have infected. “We still have a big problem collecting enough good data,” he says.

Jeremy Farrar, director of the Wellcome Trust Clinical Research Unit at the Hospital for Tropical Diseases in Ho Chi Minh City, says that much of the uncertainty over the prevalence of H5N1 could be avoided if Vietnam had better facilities for testing samples locally. “The international community continues to suggest that countries ship samples out somewhere else,” he says, “while doing absolutely nothing to invest in enhancing the scientific capacity of the Vietnamese to respond to the epidemics themselves.”

In the meantime, Tashiro adds that even if final tests confirm his negative results, “the fundamental situation has not changed”. Many are concerned that July and August will bring a new and bigger wave of flu cases in Vietnam, as happened last year during the hot rainy season. And recent events in China bode just as ill. Scientists investigating migratory birds infected with H5N1 in western China are now warning that these pose an explosive risk of spreading the virus along their migration routes as they fly south in September (see ‘Infected birds poised to take flu virus south’).

At a UN meeting on bird flu held this week in Kuala Lumpur, Malaysia, the WHO's western Pacific regional director, Shigeru Omi, warned that H5N1 remains at a “tipping point”.