Looking back on the 2008 influenza season and vaccine

Recent recommendations for universal vaccines and a change in the vaccine strains promise an interesting season next year.

This season’s influenza epidemic is now clearly on the downward curve.

At this writing, all the indices of influenza activity used by CDC are declining, save for pneumonia and influenza mortality data, which are still at peak levels. Since the mortality peak is a “late” indicator, occurring usually two to three weeks after clinical influenza activity, expect the mortality indicator to decline significantly in the several weeks ahead.

It was an unusual influenza outbreak in several respects. First, all three influenza strains were circulating and causing disease. In the early part of the season, November and December 2007, the A/H1N1 strain seemed to predominate; as the epidemic progressed, the predominant influenza A virus changed gradually over to the H3N2 strain. Influenza B was also active throughout the season, and appeared to cause perhaps 15% to 20% of all cases that occurred, at least as judged by the frequency of virus isolation.

A second unusual aspect of this season was its intensity. During a four to five-week period spanning February and the first part of March, almost all states reported widespread outbreaks, and there were many reports of serious overcrowding in emergency rooms and urgent care centers. It did not take long to back up local health care systems, reminding us once again that our health care system has virtually no excess capacity, and it does not take much of an increase in demand to overwhelm the system. A pandemic of moderate severity, such as we experienced in 1957-1958 and again in 1968, would likely collapse our present health care system. One shudders to think what a 1917-1918 severity pandemic might do.

The pneumonia and influenza mortality indicator, after it finally subsides, will probably be the most severe mortality experience in the last decade.

A third unusual feature of this year’s epidemic was the rather widely publicized vaccine strain mismatches. (In the interest of complete disclosure, I have been a consultant and voting member of the FDA’s Vaccines Advisory Committee in the influenza vaccine strain selection process for more years than I care to specify!) Yes, these mismatches probably contributed to the severity of this year’s epidemic, but we still do not have a good handle on how much.

There were two strain mismatches. The H3N2 vaccine strain was A/Wisconsin/67/2005-like, but only 17% of the H3N2 strains that were antigenically characterized were of that subtype. The great majority of H3N2 strains were a new variant, A/Brisbane/10/2007-like, a strain that evolved from A/Wisconsin/67/2005-like viruses, but antigenically distinct from them. The presence of these new variants was known in February 2007 at the strain selection meeting, but it was not clear at that time how widely they would spread; furthermore, no candidate strain for vaccine use was available. Consequently, the panel recommended the Wisconsin strain partly as a default strain, recognizing further that there would be some heterotypic immunity. The level of cross-protection achieved is not yet known but is under investigation by CDC.

The B strain mismatch is a totally different problem. Influenza B epidemiology is complicated by the fact that there are two distinct lineages of influenza B that circulate, known as the Victoria lineage and the Yamagata lineage. These two distinct lineages have co-circulated for about 10 to 12 years or more, with one or the other lineage usually predominating in various parts of the world. Importantly, there is little cross-protection from strains of one lineage to the other. Usually, but not always, strains of one lineage will predominate in a particular geographic area for a several years, but the duration of this predominance is totally unpredictable, or at least so it seems. In February 2007, the strain selection panel’s recommendation of an influenza B component from the B/Victoria lineage proved to be completely wrong. Ninety-five percent of B viruses that were isolated and antigenically characterized from this year’s epidemic belonged to the B/Yamagata lineage and were B/Florida/04/2006-like. Thus, this year’s vaccine likely provided little to no protection against influenza B.

This experience led to renewed discussions about possibly adding a second influenza B component to the vaccine so that both Victoria and Yamagata lineages would be represented. The trade-off has always been a four-component 60 mcg hemagglutinin vaccine at a reduced number of available doses vs. the standard three-component 45 mcg hemagglutinin vaccine available at the maximum number of doses, because the total U.S. manufacturing capacity is only so many doses of hemagglutinen.

As the U.S. manufacturing capacity has substantially increased in recent years, and foreign manufacturers are now licensed to market influenza vaccines in this country, the total manufacturing capacity is much less of a limiting factor than it was formerly.

Perhaps even more relevant is the recent recommendation of the Advisory Committee on Immunization Practices to broaden the recommendation of influenza vaccine to include healthy children up to 18 years of age. Since influenza B morbidity is especially prominent among children, enhanced coverage of influenza B is especially important. Thus, pressure to develop a four-component vaccine will likely be substantial in the future.

Faced with the apparent consequences of vaccine mismatches this year, the strain selection panel recommended what amounts to a clean sweep, and recommended a change in all three component strains in the 2008-09 vaccine. The H1N1 strain selected is A/Brisbane/59/2007-like; while not a problem in the now-subsiding epidemic, this represents simply an updating of the H1N1 component strain. The H3N2 strain is recommended to be A/Brisbane/10/2007-like, and the B component will be B/Florida/4/2006-like, a strain from the currently prevalent Yamagata lineage.

This was the first time in anyone’s memory that all three vaccine strain components were recommended to change. Manufacturer’s experience with growing the various candidate viruses can be unpredictable and tricky, and manufacturers prefer as much lead time as possible to gain experience in growing these viruses. The manufacturers were understandably not pleased by this recommendation but obviously had little choice in the matter. In truth, however, those manufacturers that produce vaccines for use in the southern hemisphere are already gaining experience with the recommended H3N2 and B component viruses, producing vaccines that are coming to the marketplace right now.

Although no manufacturer has yet publicly reported problems or delays, don’t be surprised if vaccine for the 2008-2009 season arrives somewhat later than usual this fall.

It has been an interesting year in influenza research as well, including both seasonal influenza as well as avian influenza. I’ll address some of those issues in the months ahead.