By Jeff Hansen - The Birmingham News al.com
UAB pharmicoepidemiologist Tim Beukelman and his colleagues combed through millions of records to study TNF-blockers. (The Birmingham News/Joe Songer)
BIRMINGHAM, Alabama -- Dr. Tim Beukelman is a medical detective -- though he and his fellow detectives might prefer the more professional term, pharmicoepidemiologists.
Their work is a kind of medical research that can be just as necessary as experiments in the lab, or testing new drugs in the clinic with volunteer patients, but it takes place way out of the limelight.
For a recent discovery — working in the world of computer databases and statistics — UAB’s Beukelman and colleagues began with a computer-choking load of billions of Medicaid billing records from 30 million children.
Somewhere in those countless files, they believed, was an answer to this important question — does a life-altering drug for juvenile arthritis, called a TNF-blocker, also have the side effect of causing cancer?
But first, some background of what was at stake.
Beukelman first encountered severe juvenile arthritis as a third-year medical student in the hospital at Washington University — in a very sick girl about 10 years old.
“She was slowly dying,” he said. “She had fever for a month and was bleeding from her lungs. No one could figure out why. ... But a rheumatologist figured it out, and saved her life.”
Juvenile arthritis is not common. It strikes children about as often as childhood cancer. Severe cases used to be incapacitating.
“I couldn’t imagine being a rheumatologist 25 years ago,” said Beukelman. “Many children were permanently crippled, disabled and in wheelchairs.”
That almost never happens today, first because of a drug called methotrexate, and then because of the powerhouse drug known as a TNF-blocker or TNF-inhibitor.
TNF-blocker interferes with tumor necrosis factor, one of the body’s molecules that acts like an alarm signal to perpetuate inflammation. In juvenile arthritis, that inflammation attacks the child’s joints.
TNF-blocker, Beukelman said, “is highly effective for many patients, and it has completely transformed their lives.”
“Their response is dramatic, and sometimes immediate,” he said. “Joint swelling, sickness and pain disappear, and the children return to normal activity. For many children with severe disease, the response is miraculous.”
Black box warning
But a shadow lay over the drug, which has been used by hundreds of thousands of patients, both adults and children, for several different inflammatory diseases.
After TNF-blocker was approved in 1998, the Food and Drug Administration began to get spontaneous reports of cancers developing in children. The FDA issued a “black box warning,” stark words surrounded by a black border on the package insert for the drug, saying lymphomas and other malignancies, some fatal, have been reported in children using the drug.
But there was a hole in that conclusion. No one knew what the rate of cancers were in children with juvenile arthritis who were not treated with the drug. There was some reason to think that that rate was higher than other children, which could mean the drug was not causing the cancers.
The answer to that question was like a needle in a haystack — hidden somewhere in the billions of Medicaid billing records.
Every state Medicaid agency collects billing data each time a patient uses the health care system — at an appointment with the doctor, a trip to the emergency room, a stay in the hospital, or a prescription at the pharmacy. The health care system sends the bill and Medicaid pays.
All of this state data is sent to the federal government, where it is combined and standardized into a huge database that medical researchers can buy. Names are removed, but each patient has a unique number.
Sifting through records from 2000 to 2005, Beukelman and colleagues found 7,812 children whose treatment codes showed they had juvenile arthritis. Some 1,484 of those had been treated with a TNF-blocker.
The bills of all of these, plus comparison groups of 974,055 children who had two different chronic diseases, asthma and attention deficit-hyperactivity disorder, were followed over time to look for evidence of cancer appearing, as shown by the doctor’s diagnosis code for cancer or prescriptions of cancer drugs.
In a paper published this week in the journal Arthritis & Rheumatism, the researchers report that having juvenile arthritis alone increased the child’s risk for cancer, compared with healthy children, though the cancer risk is still very low in absolute terms. [Read the abstract.] The FDA had not known about this higher background rate of cancer.
Furthermore — though the numbers of patients studied is low, and Beukelman is now boosting his study with four more years of Medicaid billing records — treatment with TNF-blocker or with methotrexate did not appear to increase the risk of cancer in children with juvenile arthritis.
An editorial published with the paper, titled “...Disentangling the Web,” called the paper by Beukelman a Herculean task, because low rates of both juvenile arthritis and cancer in children means that data is exceedingly scarce to find.
Besides the corresponding author, Beukelman, some of the key researchers in the paper were UAB’s Fenglong Xie and Lang Chen, who did the bulk of the computer programming, and Dr. James Lewis of the University of Pennsylvania.