Drugmaker AstraZeneca PLC's big gamble, an attempt to prove its top-selling drug works better than rival cholesterol blockbuster Lipitor, appears to have backfired.

A study meant to show AstraZeneca's cholesterol drug Crestor prevents plaque buildup in heart arteries better than Pfizer Inc.'s Lipitor showed no clear advantage for Crestor.

Two generic versions of Lipitor, the world's top-selling drug for several years, are expected to hit the U.S. market on Nov. 30. Analysts wrote Friday that the study result will make it hard for the British drugmaker to argue patients would fare better on its Crestor than on much-cheaper generic versions of Lipitor.

AstraZeneca shares fell $1.61 or 3.5 percent, to close at $45.08. The company released preliminary study results earlier in the day.

The study, known by the acronym SATURN, followed about 1,300 high-risk patients with hardening of the arteries, which can cause strokes or heart attacks, for two years. About half got a daily 40-milligram dose of Crestor, known generically as rosuvastatin, and the rest got a daily 80-milligram dose of Lipitor, known generically as atorvastatin.

Using ultrasound probes, study investigators measured plaque levels in a segment in the coronary artery of each patient, comparing levels at the study's beginning and end. Doctors believe reducing plaque thickness can cut chances of having a heart attack or stroke.

AstraZeneca said preliminary results indicated Crestor showed a greater percentage reduction in the volume of plaque in the coronary artery — the study's main goal — but that the difference was not statistically significant, meaning it could have occurred by chance. The study did produce a statistically significant reduction in a secondary goal, reduction of total plaque volume.

"The fact that this (AstraZeneca)-funded trial failed to definitively show a benefit in favor of Crestor will add to the negative pressure that Crestor is already destined to face from the imminent launch of generic Lipitor," Bernstein Research analyst Dr. Tim Anderson wrote in a report to investors. "Payers will be able to say that (AstraZeneca) itself has shown in a true head-to-head study, Crestor and Lipitor do about the same thing."

Anderson noted prior studies showing Crestor's benefit only tested it alone or compared it to a dummy pill, rather than another cholesterol medicine.

Jefferies & Co. analyst Jeffrey Holford wrote that the data didn't provide a clear positive result that AstraZeneca could have used to continue arguing Crestor is superior to Lipitor.

Still, he wrote, the positive trend and statistical significance achieved on the secondary goal wasn't a clear negative for Crestor. Holford added that the arrival of generic Lipitor will likely continue Crestor's trend of slow U.S. prescription growth.

Crestor and Lipitor are part of the drug class called statins, which lower levels of LDL or "bad" cholesterol and slightly raise levels of HDL or "good" cholesterol in the blood. Statins and other types of cholesterol drugs together comprise the second-biggest selling class of drugs in the world by revenue, after cancer medicines, and have been very lucrative for their makers.

Crestor, which came on the market well after Lipitor, brought AstraZeneca about $5.7 billion in sales last year, barely half the $10.7 billion that Lipitor made for New York-based Pfizer.

AstraZeneca will present detailed results of the study on Nov. 15 at the American Heart Association conference.

AstraZeneca and Pfizer both face significant revenue losses for several of their drugs to generic competition over the next several years.

Anderson noted AstraZeneca is expected to get generic competition in the U.S. to its top three drugs by 2016 — antipsychotic drug Seroquel in 2012, heartburn treatment Nexium in 2014 and Crestor in 2016. Together, the three account for about half the company's $33.3 billion in annual revenue, which Anderson now forecasts will decline to about $26.4 billion in 2015. He's also reduced his earnings-per-share forecast for the company by 3 percent each in 2013 and 2014 and by 4 percent in 2015.