February 2010
Feature Story

Anne Paxton

Raymond P. Podzorski, PhD, first heard about 2009 H1N1, the novel swine-origin subtype of influenza A, on a Friday in April. “There was something in the press about this novel strain being noted down in Mexico,” says Dr. Podzorski, a clinical microbiologist with ProHealth Laboratories in Waukesha, Wis. “And I thought, ‘I wonder if this could be something. Maybe we should order more influenza collection supplies.’”

But by Monday, when they went to order nasopharyngeal collection swabs and transport devices, the vendor’s inventory was already gone. “The big labs had already cleaned them out,” he says.

It was a straw in the wind—an early hint of the crisis as the H1N1 flu virus took hold and spread during 2009. Amid growing public alarm and fears of massive fatalities, there were surging numbers of cases, a scurry to authorize new tests, and a rush to deliver vaccines. Then the numbers indicating flu activity and transmission began to drop, in some cases dramatically, and the first U.S. pandemic in 40 years appeared to be over. On Dec. 4, the World Health Organization announced the global pandemic had peaked and was in decline. The CDC reported that only 3.7 percent of people tested across the U.S. and its territories were positive for seasonal flu in the first two weeks of 2010.

While it’s too early to say the 2009–2010 flu season is over, many experts believe it’s not too soon for clinical laboratories to draw a few lessons about diagnostic testing during such emergencies.

One thing was paramount: 2009 was not at all a typical flu season. Much about the 2009 H1N1 pandemic couldn’t be predicted. Beginning with the fact that the virus originated in North America rather than another continent, the pandemic threw one curveball after another—and they hit laboratories at rapid-fire pace.

“Swine flu sort of blew the roof off in terms of flu testing. It was a different animal from the usual seasonal flu,” says Karen Kaul, MD, PhD, board of directors chair of molecular pathol­ogy and director of the Molecular Pathology Division, NorthShore University HealthSystem, Chicago. “And the testing volumes were sky high.”

The usual pattern for seasonal flu is a rise in test volumes beginning in December and lasting about four months, she says. With H1N1, “It began near the end of our typical 2009 flu season. We had two peaks, with early June having the highest intensity; one day we had 70 percent positive tests, while during a typical flu season the percent positive would be around 25 percent.” There was a slow but steady stream of samples all summer; then the rates picked up again in October, finally dropping in December. Now, “we’re seeing a very little bit of seasonal flu, but that hasn’t begun to rise as it normally would this time of year.”

The virus spread geographically, probably depending on patterns of exposure and travel, Dr. Kaul notes. “We were certainly hard hit in Chicago, and there were waves elsewhere in the country at different times. Early on, it was in Texas, the east coast, and California, then the Midwest. The South didn’t get hit hard until a bit later.”

Memphis, Tenn., was one of the areas where the virus didn’t first peak until late August. “Right after school started, we observed a huge increase in the numbers of children presenting to the ER,” says Anami Patel, PhD, technical director of the molecular diagnostics laboratory of Le Bonheur Children’s Medical Center in Memphis. “We’d usually see around 150 to 160 patients each day, but by Sept. 8, our highest peak, we had 417 patients. More than 100 children were admitted to the hospital, 14 of those were admitted to the ICU, and one child died with confirmed H1N1.”

In the Milwaukee, Wis., region, the first peak of the pandemic occurred in mid-June, says ProHealth’s Dr. Podzorski. Before that, and through the summer, ProHealth was not sending samples for subtyping; every influenza A it found was what became the H1N1 strain. The laboratory started using Gen-Probe’s Prodesse subtyping reagents in August, using specimens provided by the Milwaukee Public Health Laboratory to evaluate the assay, but it did not offer the subtyping as a laboratory test until Oct. 29, after the assay received its FDA emergency use authorization.

What particularly struck him was how quickly the demand for testing can surge. “You can be doing a few tests one day, then doing hundreds the next day as word gets out to the population and they’re looking to be tested.”

ProHealth’s second wave, the week of Oct. 18, was even bigger. “At that time we were doing 150 influenza A PCR tests a day and 50 percent of them were positive—all for 2009 H1N1,” Dr. Podzorski says. The lab was subtyping at that point to see if any seasonal influenza would sneak in on this second wave in October, but it didn’t. “We restricted our subtyping to people who were hospitalized or who had individuals at home who were compromised in some way and might be particularly at risk for severe disease.”

Contrary to some expectations, the 2009 H1N1 strain was a relatively mild illness, he notes. “But in my career I had never done influenza testing in the summer. Usually it would be gone by early April, and then you wouldn’t do any until late October of the next fall. That was totally unique to have 60 tests a day in June and 30 positive for influenza A.”

In some parts of the world the number of overall cases so far has not been out of line with a normal flu season. “In Australia, they actually had more flu patients hospitalized two seasons ago. So it may turn out not to be eye-poppingly more than a traditional flu season here either,” Dr. Podzorski says. “I don’t have a crystal ball, but certainly right now as far as demand for influenza testing, we’re as low as we’ve been since the bottom end of the first peak.”

NorthShore’s molecular pathology lab, which performs nearly 100,000 tests annually, had been doing molecular testing for influenza for several years, Dr. Kaul says. “About a year ago our infectious disease doctors became increasingly concerned about resistance to Tamiflu, and they asked if we could develop a test that would discriminate various types of influenza A to help determine appropriate treatment. Seasonal H1 has become largely resistant to Tamiflu, whereas seasonal H3 remains sensitive, so viral subtyping can guide antiviral use.” The laboratory had validated an influenza A subtyping assay for this purpose that used PCR with melt curve analysis. “Then on April 22, we got a unique result: It wasn’t H1 or H3, and we didn’t know what it was. But it turned out to be one of the early swine flu cases in the Chicago area.”

“This occurred a couple of days before the big media blitz about swine flu. The following Tuesday was when the Centers for Disease Control released the sequence for the novel H1N1, and the state public health laboratory started reporting suspected novel H1N1 cases. It was another week or more before a specific assay for swine flu became available from the public health labs.”

Understandably, the initial public health response to the pandemic did not go completely smoothly. “There was quite a bit of confusion in terms of what to do,” Dr. Kaul says—partly because everything happened so fast. “The recommendations for testing varied over time,” she says, “from testing widely to testing only inpatients. We got a lot of testing during periods of time when the CDC was generally saying ‘don’t test’—in part because in many places students and employees with respiratory symptoms were required to provide a negative test result to return to school or work. And many people simply wanted to know what they were afflicted with. We were fortunate to have a validated molecular test but were really pushed to keep up with the burden of testing.”

“Ideally,” Dr. Kaul says, “rapid and definitive identification of people infected with pandemic flu could lead to isolation to limit the spread of disease, appropriate treatment, and prophylaxis with antiviral agents. And if the symptoms are not caused by flu, antiflu medication is not needed.” Conversely, if flu or another virus is identified positively in an ill patient, that may lead to a reduction in the unnecessary use of antibiotics. “It is known that during the respiratory season, patients are often over-treated with antibiotics, which has secondary effects on cost and driving up antibiotic resistance,” she says.

However, there were major variations in testing and treatment practices from institution to institution, even within the same region, in part because the swine flu pandemic exploded so quickly, leaving many institutions unable to test because of the lack of available assays. “Some hospitals in Chicago were giving everyone with respiratory symptoms Tamiflu with no testing; others were doing nothing at all, while others did a great deal of testing,” Dr. Kaul says. “In our experience, during the early weeks of the outbreak, there was still a fair amount of seasonal influenza A going around, so detailed subtyping was helpful to clinicians in managing their patients. We also have a patient population that wanted to be tested. And when the samples were coming in the door, we certainly couldn’t say no.”

Neither could ProHealth. “If they would sit and stomp their foot and demand to be tested, then it becomes a patient satisfaction issue,” Dr. Podzorski agrees. “Then we’re going to test them. You don’t want to leave them with a bad taste. The airwaves are full of all the medical centers around here advertising.”

At Le Bonheur in Memphis, Dr. Patel says, all specimens submitted for influenza testing were tested with the rapid antigen test, and if they tested negative they were routinely submitted for real-time PCR, which detected influenza A and swine flu.

“We had so many patients coming in that hospital administrators set up a tent outside our ER. From Sept. 1–18, we tested about 1,500 specimens and the prevalence of influenza was 57 percent, with about 850 out of the 1,500 being positive. There were 42 percent false-negatives with the rapid influenza test. For a while, based on CDC recommendations, if the patient was presenting with flu symptoms, we didn’t bother with testing. We’d just give them treatment and send them home.”

Dr. Patel feels it was understandable that CDC did not recommend that all patients be tested for the H1N1 subtype. “From April to the end of August, we were sending any case, either rapid or PCR, to our state health lab for a specific diagnosis, but they were expecting a large number, and there were only a couple of labs in the whole state able to differentiate.” At the same time, it did create confusion among the hospital’s physicians as to what to do. “So our stand as a hospital was: We are ready to help you differentiate.”

Randall T. Hayden, MD, director of clinical and molecular microbiology for St. Jude Children’s Research Hospital in Memphis, says his facility recently started using the molecular assays that the FDA has cleared under emergency use authorizations, but earlier there wasn’t a clear need for them. “Like most centers, we were at least initially forwarding some of our first cases to state and public health labs for subtyping. Then at a certain point, the public health authorities thought that was no longer necessary, and particularly since pretty much every case being seen was of H1N1 subtype, it was sort of the presumptive diagnosis based on influenza A.”

Dr. Podzorski feels the public health authorities managed as best they could. The state lab in Wisconsin, which offered subtype testing at no charge, got swamped, he says. “At the peak in mid-June, the turnaround time could reach 48 to 72 hours because they were overwhelmed. I would go to the government influenza Web sites every day to see what was new and what changed, and it did get frustrating because the recommendations would be different from week to week for several weeks in a row. But looking back on it, what other recourse did you have? This was the first pandemic in quite a while.”

In Dr. Hayden’s view, the CDC has done a good job of tracking and monitoring, it was quick to institute lab capacity at the public health labs, and it rapidly obtained FDA clearance for H1N1 detection reagents for PCR detection. “I think the public health labs were reasonably prepared for the epidemic. But the reality of it is that when PCR tests are sent to a reference lab, even a state or county lab, and take days or weeks to come back, they are really of epidemiologic value but hardly medical value.” Tests have much more medical value if results are available while patients are still there and treatment and isolation decisions can be made on the spot, he adds.

That NorthShore had its own PCR assays, piggybacked one off the other, was what kept the lab from being overwhelmed, Dr. Kaul says. “Quite honestly, if we had had to do it by culture or any other assay, we would not have been able to do it. Once we got the result for one assay, we could just re­flexively go into our second assay and do all our positives. It was about as smooth as we could possibly make it, and we were still really strapped keeping up with it.”

Despite the element of panic, all the publicity about swine flu was helpful, Dr. Patel believes. “They moved quickly to come up with a specific diagnostic method, and it’s very unusual that the FDA issued assay approvals through the EUA. I have never seen anything like that before, and that was a key factor. As a lab person I believed in what I was reporting out.”

Interestingly, fears that many health care workers might contract swine flu turned out to be groundless, Dr. Patel says. “In the beginning we were so worried somebody would get sick, but it turned out that the infectious disease physicians were right: The highest-risk people are those with children going to school, because those who are working with patients are taking un­iversal lab precautions. They’re washing their hands, putting on gowns and masks.” In fact, in his lab no one became ill.

He doesn’t know why the number of cases peaked in September, October, and early November, then dropped. “I have a feeling it was because of the vaccine and people’s awareness about washing their hands, and it got very cold around Thanksgiving and that could have slowed the transmission of the virus. There were probably multiple factors behind why it has gone down.”

Could the H1N1 pandemic, in the end, prove to be less severe in terms of morbidity and mortality than the average flu season? “In an average flu season, we may experience 36,000 deaths and several hundred thousand hospitalizations,” says CDC spokesman Tom Skinner. “While we may not necessarily reach those numbers with this flu season, if you look at the numbers of young people affected by this flu, compared to previous seasons, it was staggering. This has been a very, very severe flu season with hospitalizations and mortality for young people. It’s not necessarily how many were affected by this but who was affected—and it’s had a severe impact on young people and kids.”

The H1N1 pandemic presented most labs with logistical challenges, including enormous staffing difficulties. “We basically shut down everything that wasn’t necessary,” Dr. Kaul says. “Other assays that we were developing and validating—that all came to a screeching halt. We pulled extraneous staff who do order entry and clerical functions from other labs. People from all around us came to our rescue to pitch in with nonmolecular things.” Now those staff have returned for the most part to their original posts. The hospital also allowed the lab to purchase another instrument to prep nucleic acids and a higher-volume PCR instrument to process them. “That was absolutely in response to the pandemic.”

Still, the pandemic took a big personal toll on staff. “People worked many, many hours of overtime, right through the weekends, and were exhausted.” The hospital tried to find ways to reward them, sometimes intangible ways, but that was an expense that hadn’t been planned for.

Changing staffing levels can be difficult, Dr. Hayden points out. “You don’t want to react too quickly. What we did was basically follow our pre-prepared plans in increasing reagents on hand and increasing our ability to deal with high volumes of testing off-season. So, yes, there was some change in how we went about our business, and certainly some increases in resources we’ve had to expend on reagents and kits, but not to the point where it interfered with regular patient care. We didn’t compromise anything we were normally doing. We had to buy more kits, and had to put in longer hours to keep up, but aside from that, we were able to keep things going pretty much at the same level.”

Dr. Patel’s lab at Le Bonheur, which is the core lab for the entire Methodist Healthcare system in Memphis, had to adjust rapidly. “We do about 6,000 to 7,000 reportable infectious disease tests a year, and in 2009 we did about 17,000 tests. So we did double our staffing levels, and we had to double our instrumentation as well.”

At ProHealth Labs in the Milwaukee area, “we probably literally tripled the people that were involved with flu testing,” Dr. Podzorski says. “We had to get all kinds of people involved, just to get people to call all the positives. We’d have 60 positives a day and we’re just a medium-sized facility. Think of large facilities having hundreds and hundreds a day.” ProHealth took staff from other areas and has since returned them. “But we keep telling them ‘Keep your powder dry’ because typically seasonal flu won’t be peaking for another month. Who knows whether there will be a third go-round or not?”

The budget implications of the pandemic have yet to be fully faced, Dr. Podzorski adds. “Here’s hundreds of thousands of dollars every week that the state lab was burning through that were never budgeted. At ProHealth, we were spending three times what we had budgeted for flu testing, and I’m sure everyone is in the same situation.”

Unexpectedly, at least some insurers said they considered the lab’s flu testing to be experimental and would not reimburse for it, Dr. Kaul says. “We were amazed. We had some negotiations with payers, saying this is mainstream now, PCR testing is not experimental, these are FDA-approved assays with CPT codes, and at least one backed down. This was a major payer, and it happened to be the one that covers our hospital staff, but I’m sure some others out there aren’t paying.” That means patients in some cases might be receiving a bill for a couple hundred dollars, she adds.

Given the suddenness and scope of the 2009 H1N1 pandemic, these experts agree, the public health authorities did a credible job despite public pressure to the point of panic. Says Dr. Kaul, “I’m sure there was an element of panic as far as public policy, but you know, we had no way to predict where this was going to go.” Dr. Hayden adds: “I’d rather there be an over-reaction than an under-reaction. That’s when an epidemic can really have terrible effects.”

Top of mind for every lab is what will happen in the coming year. It’s possible that more people than usual got the seasonal vaccine in October last year, Dr. Kaul notes. “It was out earlier and worked well, better than many other years, and perhaps there was more of a public rush to get that.”

In theory, if the exact same virus returns next flu season, the public should largely be immune to it, so it will be a question of how much the virus changes, she says. “H1N1 was a recombination. There was a reassortment and a recombining of viral fragments, probably in animal hosts. It was immunologically quite a bit different from normal seasonal flu, and that’s why it swept through. Basically no one was prepared to fight it off because we hadn’t seen anything like it before. That’s why periodic pandemics can be so devastating.”

Every few years viruses shift, which means a bigger series of changes that often involves picking up stray bits of nucleic acid, Dr. Kaul explains. “All of these viruses are sneaky and they change sequences. Depending on where you targeted probes and primers and the like, you can be fooled,” she notes. For example, her lab has been seeing sequence variants for the past few months. “We can still detect the virus and call it influenza A and likely swine flu, but they appear different. And there are other assays out there in commercial use that can’t see the variant virus anymore.”

St. Jude’s had a fairly extensive plan for pandemic flu in place before the H1N1 virus hit, Dr. Hayden says. “But of course nothing tests a pandemic plan like having a pandemic.” Although things went pretty much according to the plan and there were no major stumbles, “there are always things that one might want to redo. Our plan is to go back at the end of the flu season and sit down and do a thorough assessment. From our standpoint, this was not as severe as it might have been, fortunately, so I think if another wave hits us tomorrow, we’ll be prepared because we’ve ramped up our capacity in anticipation of perhaps higher volume than we ended up getting.”

For a small lab, he notes, it does become difficult to face a logarithmic unit increase in testing. “You don’t have as many people to move around and cover. As we move to higher-throughput tests like the molecular amplification tests, then the playing field becomes leveled a bit more. Having automated methods in place that minimize the direct linear relationship between volume and staffing requirements is what’s going to help us in the future in meeting these types of test challenges.”

The outbreak made clear the importance of laboratory-developed tests like the PCR tests at NorthShore, Dr. Kaul believes. “Labs like ours played a key role here. The public health labs are not the frontline for diagnosing and managing patients. The Illinois Department of Public Health was so swamped it couldn’t handle the volume. They are not staffed enough and able to turn results around quickly enough to handle patients in medical situations. They can track outbreaks. They can give you results a week later, but if you’re in the trenches dealing with patients, they cannot turn around fast enough.”

“There’s a lot of scrutiny of infectious disease testing and how it should be overseen by CAP and CLIA and the federal government, but if we had not had this assay, we would have had nothing,” Dr. Kaul continues. “And it just happened out of luck, because we had developed this test for another purpose.” Another area that needs to be explored, she believes, is how hospital labs can get results more quickly to physicians who are caring for the patients.

Even though the pandemic may appear to be winding down, neither medical facilities nor clinical laboratories can afford to relax, Dr. Hayden warns. “What’s going to be interesting is whether we’re taking a big sigh of relief and it’s going to come back, or whether there won’t even be a normal flu season this year. If we did get another wave and places were even harder hit, how would they deal with it, having already used up their regular budget?”

Whatever occurs from here on out, the pandemic does underscore the importance of emergency preparedness planning, Dr. Podzorski emphasizes. “These plans you make in advance really are worth it. A few years ago, everybody at ProHealth Care became concerned about avian influenza because of the high mortality associated with it, and there was planning on how to handle the influx of patients. That certainly came in handy with the H1N1 pandemic.”

Related Links

• A testing first: moderate complexity PCR for H1N1