The arsenal of antibiotics strong enough to squelch nasty bacteria is rapidly dwindling

worldwide, which makes worried infectious-disease doctors more intent than ever that the drugs be deployed only when strictly needed.

These specialists know that every antibiotic carries its own risks, and that the more frequently and broadly a drug is used, the more likely it is that harmful microbes will develop tricks to sidestep it. But a team of researchers in the Netherlands, where a more selective use of antibiotics has led to much lower levels of resistant bacteria than are circulating in the United States, thinks the medical finger-waggers have not gone far enough.

"As doctors, we&39;ve paid a lot of attention to questions of which antibiotics we should use to treat what sorts of infections, but have focused much less on how long that treatment should last," said Dr. Jan Prins of the Academic Medical Center in Amsterdam.

In a small but provocative study published in the June 10 issue of the British medical journal BMJ, Dr. Prins and colleagues from nine hospitals suggested that even some cases of pneumonia — a potentially life-threatening disease — could be treated with a three-day course of antibiotics, rather than the conventional 7- to 10-day treatment.

The Dutch study analyzed the cure rates of 186 adults who had been hospitalized with mild to moderately severe pneumonia. All received three days of intravenous amoxicillin to start. After that, the 119 who were showing substantial improvement were randomly divided into two groups; about half continued with another five-day course of oral amoxicillin, and the others got look-alike sugar pills. Neither the patients nor the doctors knew who was getting which treatment until the end of their participation in the study.

By the end of treatment, roughly 89 percent of the patients in each group were cured of their lung infections without further intervention. In a commentary accompanying the study, Dr. John Paul, a microbiologist at Sussex County Hospital in Brighton, England, writes that, at least for a subset of patients with uncomplicated, community-acquired pneumonia, the finding "suggests that current guidelines recommending 7-10 days should be revised."

As lead investigator of the Dutch study, Dr. Prins was not ready to go quite that far. He cited the study&39;s small size and the seriousness of the illness as a reason to wait until the finding is independently replicated before advising a wholesale change in practice.