In a rebuke of the safety of Abbott Laboratories diet drug Meridia, eight of 16 members of a U.S. Food and Drug Administration advisory panel said this afternoon that the drug should be withdrawn from the U.S. market.

Meanwhile, six of the panelists said the drug should only be prescribed by “specially trained physicians” and include a strict FDA black box warning noting the new limits.

The other two panelists also said a new boxed warning should be added to alert consumers of increased risks of heart attacks and closer monitoring of patients by clinicians.

None of the panelists among the Endocrinoligic and Metabolic Drugs Advisory Committee said the drug should remain on the market with the labeling in its current form. So even if the Food and Drug Administration decides to keep it on the market, doctors will be encouraged to prescribe it in limited circumstances.

The FDA usually follows the advice of its panels but is not required to do so.

The U.S. Centers for Disease Control and Prevention estimate that two in three Americans are overweight and one in three are obese, and the drug market is hungry for an effective diet pill. But the idea has been difficult to pull off.

There has not been a prescription pill for weight loss approved since 1999, when Xenical, which works by blocking the absorption of fat, was approved.

And although the FDA allowed a version of Xenical known as Alli to be sold over the counter, doctors say gastrointestinal side effects such as diarrhea have kept many consumers from taking it for long periods.

Meridia has not been a big seller. Revenues from the pill have deteriorated amid criticism of its heart risks. Meridia’s maker, North Chicago-based Abbott, says it no longer promotes the drug in the U.S., where its sales are projected this year to be $30 million.

Prescription diet pills have had trouble winning respect with consumers and doctors because of safety issues and side effects. In 1997, a diet drug combination known as fen-phen was yanked from pharmacy shelves after it was linked to heart valve damage.

Such problems with past treatments have led drug developers to turn in recent years to combinations of drugs already on the market, thinking years of safety and efficacy data on low-dosage versions would help win approval.

But that strategy did not work for Vivus Inc., which in July had its anti-obesity drug Qnexa rejected by an FDA advisory panel by a 10-6 vote, in part due to heart-attack risks.

Following the panel’s decision on Qnexa, Biotech Stock Research, a research firm that follows drug development, Tweeted that it believes “no drug will ever be proven to be as safe as exercise in the minds of reviewers.”

Analysts see a huge category if an effective diet drug is ever approved, especially since the obese tend to already have chronic conditions such as diabetes, high cholesterol and hypertension — conditions that generate some of the biggest categories of drug sales in the U.S.

Cholesterol drugs, for example, are the second most lucrative sales category, tallying $14.3 billion in revenue last year, according to market research for IMS Health. Antipsychotics were the top-selling class, with $14.6 billion in 2009 U.S. sales, IMS said.

Lorcaserin has not shown the similar heart-attack risks as Qnexa or Meridia, so some analysts believe peak sales could one day surpass $500 million annually. It works similarly to how fen-phen worked, stimulating a receptor in the brain to help block signals linked to appetite so the person taking the drug eats less.

San Diego-based Arena Pharmaceuticals Inc. and studies have shown lorcaserin is more selective in the receptors it targets than the fen-phen combination, so its side effect profile is better, data seems to indicate.

But an FDA staff report issued Tuesday on lorcaserin in preparation for Thursday’s panel said the drug satisfied the agency’s “categorical efficacy criterion” only “by a slim margin” and raised safety issues such as memory problems and a link to heart-valve damage.