If confirmed in future studies the research, in the March 15 issue of The Journal of Infectious Diseases, may drive a major shift in strategy for controlling TB.

A major difficulty in treating patients with pulmonary TB is that the organism can become progressively resistant to standard therapy. This resistance was long thought to be acquired through mutations in the infecting strain when the treatment regimen was inadequate or the patient did not comply with it. More recently, studies of the genetic make-up of Mycobacterium tuberculosis (M. tuberculosis) strains have shown that resistance can also result from re-infection with a new strain that is already drug-resistant, sometimes against multiple drugs.

The authors of the new study, Qian Gao, PhD, and coworkers in Shanghai, China and elsewhere, used molecular genetics and drug susceptibility testing to investigate patients with TB who were treated in Shanghai hospitals during 1999-2004. They focused on 38 patients from whom samples were available before and during treatment. The researchers found that the strains of TB in the samples taken before treatment were genetically different from those taken during treatment in 87 percent (33 out of 38) of patients.

To determine the relative proportion of drug resistance caused by re-infection or mutation, the authors excluded six patients who were initially infected with resistant TB and then became drug-susceptible or resistant to fewer drugs. In the remaining 32 patients, the initial sample was drug-susceptible or resistant to at least one drug and the subsequent sample resistant to one or more drugs. Of these patients, 84 percent (27 patients) had before-and-during samples with different genetic patterns and only 16 percent (5 patients) had identical patterns. Thus, there were more than 5 times as many cases caused by re-infection compared to mutation.

"It was surprising to find a high rate of primary drug-resistant strains among treated patients," said Dr.Gao. "This overturned the common belief that drug resistance among treated patients is always acquired."

The investigators also noted that two patients in the study had multidrug-resistant strains in both their first and second sample, and that 10 others had multidrug-resistant strains in their second sample; genetic testing showed that 9 of the 10 patients had a different strain in the second sample. The most serious kind of drug-resistant disease therefore accounted for about a third of patients with drug resistance.

Limitations of the study included the exclusion of many patients without sample results, reliance on previously collected data in which some patients might have been misclassified, use of computerized drug susceptibility data, and the unknown contribution of mixed infections. Nevertheless, the findings are a warning. Although better diagnostics, drugs, and effective vaccines for TB are clearly needed, the authors said, "Our findings highlight the urgency of accelerating efforts to interrupt the transmission of drug-resistant tuberculosis." The research shows improved methods of preventing TB transmission may be needed in the very facilities and communities where TB patients are treated.

idsociety