Designers of anti-obesity drugs have suffered three major setbacks, but the potential reward from treating the world's fat epidemic is so great that their quest is unlikely to be deterred.
After investing a sum rumoured to be in the hundreds of millions of euros (dollars), Sanofi-Aventis of France announced last Wednesday it was abandoning its drug rimonabant, which had stoked huge expectations at its launch in 2006.
Rimonabant, brand name Acomplia, ran into a flurry of ever-tougher warnings from European watchdogs about potential psychiatric side effects, including depression.
Also Wednesday, the US giant Pfizer put a stop to tests of a prototype in late-stage development called CP-945,598, citing regulatory hurdles.
"The risk/benefit profile in this class of drugs was lower than expected for obtaining market authorisation," said Catherine Baulac, in charge of new products with Pfizer's French subsidiary.
And on October 2, Merck of the United States pulled the plug on its own experimental obesity fighter, taranabant, because of concerns about anxiety and depression at high doses.
The main causes of obesity, overconsumption of fatty or sugary food and a sedentary lifestyle,are well known. But the molecular machinery that drives it is in many ways obscure and, it is now suspected, more complex than thought.
Treating obesity entails a limited basket of options, from lifestyle changes to drugs and gastric-bypass surgery.
But when it comes to that middle choice, the most promising class of new prescription medicines for treating long-term obesity has just been wiped out.
The three scratched drugs belong to a group called cannabinoid receptor antagonists.
They take on the same target in the central nervous system as marijuana. But instead of sharpening appetite, the "munchies" associated with smoking cannabis, they work in reverse, dulling the urge to eat.
"There are now just two drugs left, orlistat and sibutramine," Colin Waine, a doctor who is chairman of a British organisation, the National Obesity Forum, said.
Anders Sjoedin, a specialist in obesity drugs at the University of Copenhagen, Denmark, said cannabinoid receptor antagonists had now been discredited as an approach.
"Cannabinoid receptors not only regulate the appetite, they also affect mood," he said.
Orlistat, marketed as Xenical or alli, curbs absorption of fat in the intestine by blocking a pancreatic enzyme. Sibutramine, sold as Reductil or Meridia, affects levels of a brain chemical called serotonin, believed to influence feelings of hunger or satiety.
Both are "very useful," said Waine. They come, though, with a list of side effects -- including, in orlistat's case, the risk of sudden, oily faeces.
As proof of the glittering allure of the obesity market, drug engineers are exploring unusual paths.
According to the UN's World Health Organisation (WHO), around 400 million adults were obese in 2005, and the tally is expected to balloon to more than 700 million in 2015.
In the lab, at least, are potential rivals to orlistat in the field of lipase inhibitors; new neurotransmitter inhibitors that reduce appetite; and even a hydrogel pill that expands in the stomach to give a sense of fullness.
But, should these prototypes ever get the green light to go on sale, will they meet the hopes of people desperate to lose dozens of kilos in a year, rather than just a few?
Even more important, are they safe?
Eleven years ago, the search for an obesity drug ran into controversy when fenfluramine, an appetite suppressant, was banned in the United States over fears of its effect on the heart. Last week, a study published in the journal BioMed Central found fenfluramine's damage to cardiac valves was visible seven years later.
Such questions pose a dilemma for scientists much as they do for drug regulators.
Obesity is a huge and growing disease, for which people are clamouring for a quick fix. Yet it is also a complex disease, and no drug is without side effects to some degree.
Sjoedin said that researchers believed that regulators, especially in the United States, were becoming increasingly demanding about proposed obesity drugs.
Watchdogs were wanting to see more data about adverse effects from proposed drugs, as well as further proof of benefits other than weight loss, such as lower blood pressure.
"Obese people will be around, and I'm not sure that surgery is the solution for all them," said Sjoedin. "In that respect, there is a need for drugs. But no one is going to make a drug unless they can make money out of it in the end."