H1N1 vaccine: Counting side effects

Pregnant women are considered at high risk for suffering complications or death from the new H1N1 pandemic swine flu. So they’re near the top of the list for getting vaccinated. A new international study calculates that up to 400 out of every million pregnant women who receive such swine-flu shots will experience a miscarriage within 24 hours.

But not BECAUSE of their flu shots.

These miscarriages would have normally occurred even if the moms-to-be had foregone vaccinations. And that’s because the average background miscarriage rate for any given day (among women far enough along to recognize they are pregnant) is about 397. Or so reports a new international analysis, which appeared online October 30, ahead of print, in the Lancet. To see an impact of the vaccine on pregnancy losses, miscarriage rates would have to spike well above that roughly 400-per-million figure, it notes.

If policymakers, physicians and the public don’t understand the magnitude of background rates for diseases and health impacts, they risk inappropriately attributing certain adverse events to vaccines.

Internet web browsing together with public concern about the safety of vaccines “have increasingly allowed for spurious associations to be promoted as fact,” note Steven Black of Cincinnati Children’s Hospital and his international team of colleagues. Reports of false associations “can and do disrupt immunization programmes, often to the detriment of public health,” these researchers argue in the Lancet.

For the new analysis, the team pored over published papers reporting incidence figures for a number of conditions that the U.S. Food and Drug Administration noted could turn out to be possible side effects of vaccinations. These included infectious polyneuritis, also known as Guillain-Barré syndrome (an autoimmune disease that can cause muscle weakness and sometimes paralysis); multiple sclerosis; sudden inflammation of the optic nerve; Bell’s palsy; anaphylaxis (a potentially life-threatening whole-body allergic reaction); seizures, sudden death (within two hours of symptoms developing); an autoimmunity-linked reduction in blood platelets (thrombocytopenia); miscarriages and preterm labor.

The researchers found that incidence rates for some of these conditions can be relatively high — and can vary widely by country, gender, age or other factors.

Take Guillain-Barré syndrome. Some 1976 swine-flu shots were associated with a seeming up-tick in Guillain-Barré cases, so that disorder will be on the radar screen of epidemiologists looking at the safety of the new H1N1 vaccine.

The background incidence of this syndrome in Finland ranges from about 0.2 cases per 100,000 in boys who are 17 and under (about one-tenth the rate of girls there) to 10 per 100,000 men 65 and older (a somewhat higher rate than afflicts this Scandinavian nation’s Golden Girls). In the United States and Britain, however, boys and girls have roughly comparable rates of Guillain-Barré syndrome, although they vary by country — with slightly fewer than one case per 100,000 in the United Kingdom and almost double that in the States. Those rates climb to only between 2.5 and 4.5 per 100,000 in the UK’s post-65 set and peak at just 2.3 to 3.3 per 100,000 women and men over age 65 in the United States.

Black’s team argues that it’s important to identify background national rates of this and other diseases before panic sets in as someone attributes his or her flu shot to triggering Guillain-Barré paralysis. For instance, the new analysis predicts that perhaps four cases of this autoimmune disease per 10 million vaccinated people might be expected to occur within one week of their getting flu shots — just based on background rates — and 22 cases within six weeks.

For sudden death, background incidence would suggest that among every 10 million vaccinated people, five or six such deaths might spontaneously occur — unrelated to flu — within six weeks of getting flu shots.

One problem, of course, is that if five people in the Boston metro area develop Guillain-Barré syndrome within a week of getting flu shots, someone might attribute it to the vaccine, even if no other cases occur across the rest of the country. Such apparent hot spots occasionally emerge, only to later turn out to be random flukes. Before they’re confirmed as such, however, news accounts might unleash a panic that shuts down vaccination rates in Boston and elsewhere — allowing pandemic flu to spread unchecked.

In a commentary accompanying the new Lancet analysis, Frank DeStefano and Jerome Tokars of the Centers for Disease Control and Prevention in Atlanta point to another underlying problem: How vaccine-monitoring systems collect reports of adverse events. They provide a numerator that might point to potential problems. But unless there’s also a denominator, it can still be hard to know if adverse incidents are truly unusual in number. And, in fact, most nations have not been good about collecting fairly real-time vaccination counts that are stratified by age, gender or region.

The United States and other nations will try to glean such data in coming months. But unless they’re enormously successful — and quick — DeStefano and Tokars argue that epidemiologists could find themselves seriously hampered in trying to establish whether serious vaccine side effects are emerging — or only seem to be.