OPTIONS VI: Surveillance System Monitors Flu Vaccine Efficacy in Children

TORONTO, June 21 -- To quantify flu vaccine's effectiveness for children with more specificity, a new surveillance network is able to provide consistent and potentially valuable estimates, CDC researcher said here.

Unfortunately, the system's results for the 2003-04 and 2004-05 flu seasons appear to show the trivalent inactivated vaccine wasn't very effective, according to Marika Iwane, Ph.D.

The challenges for the system in a given year include an unpredictable number of flu cases, vaccines that don't match the circulating strains, and low vaccination rates, Dr. Iwane told a session of the Options for the Control of Influenza meeting.

Despite those challenges, she said similar point estimates of vaccine effectiveness were seen in three study groups, using the same flu cases, but different controls.

Over the two seasons, Dr. Iwane said, the vaccine appeared to work more or less well if children were given the full two-dose treatment, but was less useful if they missed a dose.

However, there was variation in effectiveness from year to year and between older and younger children.

She and colleagues used the New Vaccine Surveillance Network, which was set up in 1999 to provide longitudinal data to help measure the impact of respiratory disease vaccines in children younger than five.

It includes prospectively enrolled children in three counties in the states of New York, Ohio, and Tennessee, Dr. Iwane said.

For this study, the cases were children in the network who went to hospital with a respiratory illness that was confirmed to be flu.

They were compared with three control groups:

• Children in the network who had a non-flu respiratory illness, a so-called negative control group.
• Children, drawn from representative practices in the three countries, whose charts were reviewed.
• A cohort of children found by random digit dialing. This study was only conducted in 2003-04.

In 2003-04, there were 193 cases, compared with 447 who had a non-flu illness, 4,795 in the chart-control group, and 1,272 in the random-dialing cohort.

The following year, despite efforts to ramp up surveillance, the number of confirmed flu cases fell sharply, Dr. Iwane said. "They actually enrolled fewer that year because it was a very mild season," she said.

For 2004-05, there were 84 cases, 630 with a non-flu illness, and 4,628 in the chart control group.

Comparing the groups, the researchers calculated hazard ratios and defined the vaccine effectiveness percentage to be unity minus the hazard ratio times 100.

In 2003-04, vaccine effectiveness ranged between 4% and 17% (depending on the control group) for fully vaccinated children from six months through 23 months, but the confidence intervals were wide and the results weren't significant, Dr. Iwane said.

The following year, however, vaccine effectiveness was 64% compared with the negative controls and 66% compared with the chart control group and both results were significant, she said.

For children from two through five years in the 2003-04 season, the vaccine effectiveness point estimates were high -- about 80% -- but the confidence intervals included unity, Dr. Iwane said.

For the older children the following year, results were mixed and non-significant, probably owing to the small number of cases, she said.

But the disappointing results shouldn't be taken too seriously, said vaccine expert John Treanor, M.D., of the University of Rochester (N.Y.) Medical Center.

"What you're looking at is a report on a system that may ultimately prove to be very useful," he said, rather than a mature report on the effectiveness of the vaccine.

"We have very little information on the effectiveness of vaccines in children generally," he said. "Do changes in antigen drift have an effect? How much of a problem is it? We really don't know."

"It's really a system that's in the making, rather than a definitive result."