The Food and Drug Administration Tuesday urged at least a one-third reduction in the recommended concentration of the obstetric anesthetic bupivacaine. The amount of the reduction was reported incorrectly yesterday.

The Food and Drug Administration yesterday urged at least a threefold reduction in the recommended concentrations of the most widely used obstetric anesthetic after at least 16 women died during labor.

The women were among 20 who have suffered a cardiac arrest when the local anesthetic, bupivacaine, was injected into a vein accidentally instead of into the space surrounding the spinal cord and its membranes.

In some cases, the babies being delivered died too, although the FDA could give no specific figure.

The pain killer's three manufacturers--Abbott Laboratories, Astra Pharmaceutical Products and Breon Laboratories--are cooperating, the FDA said, by mailing a joint letter to doctors and hospitals advising them to use lower doses and making other preventive recommendations.

The drug--also known as Marcaine and Sensorcaine--has been used in its high concentration in more than a million women in the past two years, in short, one woman in seven or eight in labor. Its greatest use has been in women whose babies are being delivered by Cesarean section.

It "is probably one of the safest" obstetrical anesthetics for mother and baby, said Dr. Patricia Russell, an FDA drug supervisory officer.

But occasionally, even in the best hands, the drug is injected accidentally into a vein, some doctors said yesterday. It is then that a toxic reaction, followed by cardiac arrest--meaning either heart fibrillation, then stoppage, or immediate stoppage--may happen.

Russell said that far lower concentrations than the .75 percent concentration in common use--a .25 percent concentration or even in the view of "most anesthesiologists" a .5 percent concentration--can provide enough anesthesia for either Cesarean or normal delivery. "This is a step to cut the risks of obstetrics," she said.

The action was called "a good starting point" but no more, however, by Dr. George Albright, a Stanford University anesthesiologist who began raising questions about the drug in an editorial in the journal Anesthesiology four years ago.

He said that at an FDA conference this fall, he will urge further restrictions on the pain killer's use both in obstetrics and general surgical anesthesia.

Albright, director of obstetrical anesthesia at Stanford University Medical Center, was "until recently . . . a voice crying in the wilderness" on bupivacaine's dangers, Russell wrote in an internal memo last month.

In May, 1982, an FDA advisory committee recommended against any restrictions on the drug's use, pending more study. "The picture is much clearer now than it was then," Russell said. "There are a lot more cases and there is animal data."

The advisory committee was polled, she said, and agreed on the present step.

Albright served notice that he will urge more restrictions. Some people have blamed the drug's problems on poor medical management, he said, "but I feel very strongly that it's the drug's toxicity. I feel very strongly that it essentially poisons the heart. Every once in a while, no matter what you do, you get the drug in a vein and the patient goes into seizure. Usually this is temporary, but occasionally the patient goes on to arrest."

In Britain, he said, the drug is used in obstetrics in only a .5 percent solution and is given far more slowly to obviate any ill effects.