WORST-CASE SCENARIOS don't come much worse than this. Eight years ago scientists identified a terrifying new disease that had the potential to infect up to ten million people in Britain.

Each victim would sicken in the same horrible way: his or her brain would be eaten from within by a deadly protein, mental and physical decline would be swift and end only in silent, bedridden death. Fates were already sealed - anyone who had ever eaten a burger or a meat pie was theoretically at risk.

The predicted epidemic of variant Creutzfeldt-Jakob disease (vCJD) has yet to come. Now there are whispers that it may never materialise. So can we stop worrying? "It looks that way," says Dr Azra Ghani, from the Department of Infectious Disease Epidemiology at Imperial College, which has been monitoring the disease since its inception. Ghani, who is preparing a paper on the up-to-date figures relating to vCJD, is encouraged by the fact that fewer people succumbed to the disease last year than in previous years. Seventeen people died in 2002, compared with 20 in 2001. The first deaths (a total of three) were recorded in 1995, with the number peaking at 28 in 2000.

"Statistically speaking, the figures are in decline, and the upper boundaries on our estimates are coming down," Ghani says, adding that the drop in two consecutive years is more than just a lucky dip; it is a genuine fall and could signal the beginning of the end of the biggest food scare Britain has seen so far.

The Public Health Laboratory Service, which analyses the figures every quarter, is also being upbeat. Its website notes: "There is now statistical evidence that the epidemic is no longer increasing at the rate seen previously, and that it may have reached or be reaching a peak." However, it doesn't rule out a second peak.

Professor James Ironside, the neuropathologist who heads the National CJD Surveillance Unit in Edinburgh, says he is also "cautiously optimistic" but not complacent. It is a disquieting fact that sporadic CJD, which appears spontaneously in individuals and pre-dates vCJD, is on the rise. There has been speculation that this disease, which has a different neurological hallmark from vCJD, might also have something to do with BSE, although nobody has come up with a convincing piece of evidence to connect them. It would also have to explain why sporadic CJD is found in countries free of BSE.

"I would agree that a lot of the worst-case scenarios, involving hundreds of thousands of people contracting vCJD, can be eliminated," Ironside says. "It's true that the death rate and the incidence of new cases has dropped off. But I'm not sufficiently reassured that this is the end of the story. There are still questions about individual susceptibility to be answered."

In particular, all the victims so far have shared one genetic characteristic: they have a genotype known as MM, a code for certain amino acids found in the brain. Around 40 per cent of the population share this MM genotype (the rest of us are either MV or VV). So, either 60 per cent of us struck it lucky in the genetic jackpot and are immune, or we still have the full horror of vCJD to come. Scientists do not have a clue which one is true.

Ghani and Ironside are more optimistic than Professor Stanley Prusiner, the Nobel prizewinning scientist who discovered prions. Last year Prusiner suggested that eating the muscle tissue of infected animals might be just as dangerous as consuming their brains and spinal cords. When he injected laboratory mice with scrapie, the ovine form of mad cow disease, he found worrying levels of prions in the mice's muscle tissue. The inference was that eating prime cuts of beef, which consist of mostly muscle, might be as deadly as eating tissue from the brain and spinal cord. However, in exhaustive tests when the BSE scare was at its height, scientists tried, and failed, to find prions in muscle tissue taken from BSE-infected cattle.

"Prusiner's work is interesting, but it doesn't seem to match the observations on BSE," Ironside says. "People have looked for prions in muscle and have not been able to find them, in humans or animals." That is possibly because what happens in one species doesn't always happen in another; also, injecting is a much more potent way of delivering prions than eating them.

A major study on 8,000 appendix-tissue samples carried out last year should have clarified how widespread vCJD is, but the results were mixed. Dr David Hilton, a neuropathologist at Derriford Hospital in Plymouth, discovered one case of vCJD. It was one more than anybody wanted to find, but, as researchers are aware, it is almost impossible to extrapolate from that one positive.

"One positive in 8,000 equates to roughly 120 cases per million, but I would say that no firm conclusions can be drawn," Hilton says. Ghani, who collaborated on the study, is more forthright. "It's a higher rate than we would have expected, and is a slight inconsistency, which means that we can't totally write this off." Scientists now plan to collect and scan 50,000 samples of tonsil tissue, which is known to be prion-rich in infected people. It will be at least two years before the results are known.

Meanwhile, clues about whether the vCJD scare is over lie in understanding the incubation period. BSE in cattle was first described in 1986, and the first deaths from vCJD happened in 1995. The 129 cases so far - 121 dead and eight still living - might be as bad as it gets. On the other hand, perhaps the victims so far had unusually short incubation periods. That means there might be many more people who are infected but, for genetic reasons, the symptoms will take longer to emerge. There is already a precedent for this - cases of kuru, a fatal neurological disease associated with cannabilistic rituals in Papua New Guinea, are still being diagnosed today, even though tribes stopped practising cannibalism in the Fifties.

But the longer the incubation period, the less relevant vCJD will become. This is because, as time goes on, it becomes more likely that an infected person will die from something else.

This, Ironside insists, makes it imperative to carry on research into the genetic factors that contribute to the development of vCJD. In particular, why do two-thirds of us appear to be untroubled by the disease? Or is there another vCJD timebomb waiting in the wings, just ticking more slowly?