In a new report, researchers say they might have found some signs of the former, but they also acknowledge that the evidence is extremely weak.

"The idea of whether treatment for ADHD affects the risk for other conditions is a really interesting thing and there is a lot of back and forth," said Dr. Paul Hammerness of Harvard Medical School in Boston, who led the research.

About 2.7 million children and teens in the U.S. take stimulants for ADHD, such as Novartis' Ritalin, Shire's Adderall and Janssen's Concerta.

Studies in rats have shown that giving the animals stimulants might whet their appetite for cocaine later on. But there have been mixed findings on the link between stimulants, substance abuse and smoking in humans.

In the new study, which was funded by Janssen, Hammerness and colleagues put 154 adolescents with ADHD on Concerta for up to two years. After an average of 10 months of treatment, about 7 percent of the teens or pre-teens reported smoking.

The rate was lower than the percentage - about 20 percent - in a comparison group made up by adolescents with ADHD who were not taking stimulants when asked about their smoking habits.

But that group came from earlier studies and might not be comparable to the kids Hammerness and his colleagues studied. In fact, when the researchers accounted for differences is behavioral problems between the two groups, the gap in smoking rates disappeared.

Hammerness, who has ties to several manufacturers of ADHD medications, said no firm conclusions could be drawn from his study, which was published in The Journal of Pediatrics.

"You can't conclude anything from the study, nope," he told Reuters Health. "This is the nature of the beast."

But he also said that placebo-controlled studies - the gold standard - are hard to do because not treating children with ADHD for extended periods might be considered unethical.

SOURCE: bit.ly/OAyMKd The Journal of Pediatrics, online August 9, 2012.

• A Permanent Economic Emergency