Chapter 11. Influenza and Antibiotic Resistance

In earlier chapters, we focused on diseases that tend to be problems in specific geographic regions. We now turn to a viral disease, influenza, which spreads so rapidly that it can affect the lives of almost everyone on the planet. Here, we briefly describe influenza biology and discuss how resistance may neutralize a major portion of our preparation for pandemic flu. Three general situations are considered: seasonal flu, human flu pandemics, and avian flu.

Seasonal Influenza Virus Is Controlled by Vaccines

Seasonal influenza, which moves around the world every year, is blocked by vaccines. Seasonal flu is thought to begin in wild waterfowl in Asia. From there, it spreads through domestic poultry and eventually reaches pigs. Viral reassortment occurring in pig tissues infected with both bird flu virus and human flu virus can produce a new version of influenza that is highly infectious to humans. (The viral genetic material comes as eight separate segments; when a cell is infected with two virus variants, progeny viruses acquire their eight genetic segments as a mixture, some derived from one parental virus and some from the other.) After the virus starts to spread among humans, it is isolated and used to prepare vaccine that is administered in the fall. Because the majority of infections occur months later in late winter (February and March in the Northern Hemisphere), there is time to prepare and distribute the vaccine. Unfortunately, new vaccines need to be prepared each year because the major circulating virus is different each year.

Antiviral Resistance Has Emerged Among Seasonal Influenza Virus

Flu vaccines are only 70–90% effective.²²⁹ Consequently, antiviral drugs have been developed for persons unable to develop a good immune response to the vaccine. Influenza viruses fall largely into two general types called A and B. (In the 2008–09 season, 77% were type A, 23% were type B.²³⁰) Two antibiotic classes have been used for influenza A. The adamantanes (amantadine, rimantadine) inhibit influenza virus membrane protein-2 (M2). This protein is part of a channel required for passage of protons that help uncoat the virus. The adamantanes prevent flu virus from taking over the host cell. These drugs, which have been available since 1964, are ineffective with influenza B because this virus lacks the M2 protein.²³⁰ The other antibiotic type, the neuraminidase inhibitors (zanamivir, oseltamivir [Tamiflu]), blocks the activity of the viral enzyme that breaks down glycolipids and glycoproteins. These agents prevent release of progeny virus from infected human cells. Neuraminidase inhibitors were designed for influenza A; influenza B tends to exhibit lower susceptibility. Zanamivir and oseltamivir target different sites on neuraminidase; consequently, the two drugs show little cross-resistance.²³¹ Oseltamivir is taken orally and has been applied throughout the world; zanamivir has been used largely in Japan.

In principle, antibiotic resistance emerges as a direct result of selection pressure acting on spontaneous mutants. Resistance can then be maintained in the virus population either by continued antibiotic pressure due to treatment or by the resistance mutation being located near other viral mutations that confer an advantage to the virus.²³² (Two mutations that are close to each other in a viral RNA segment tend to re-assort together and maintain a tight association.) In the latter case, resistant virus can spread to many persons who never receive antibiotic treatment.

In 2002–03, the circulating strains of seasonal influenza began to display adamantane resistance in Asia, perhaps stimulated by treatment of domestic birds.^230,232 The emergence of resistance was exacerbated when the agents were added to over-the-counter cold remedies in Russia and China.²³² By 2005–06 amantidine resistance became extensive in the United States,²³² and by 2008 the viruses displayed such a high prevalence of resistance that the compounds were no longer recommended for treatment.²³²

Resistance to the neuraminidase inhibitors has also been appearing, even though these drugs have not been used extensively for domestic birds.²³⁰ With clinical trials of oseltamivir, resistance was reported in 2% of treated patients (18% in treated children).²³³ Such high numbers indicated that resistance would arise readily if the drug were widely used. However, prior to 2007 oseltamivir in untreated patients was rare (less than 0.3%).

In the 2007–08 season, the worldwide prevalence of oseltamivir resistance rose to 15% among seasonal influenza A H1N1 isolates.²³¹ (Subtypes are defined by differences in two viral proteins, hemagglutinin and neuraminidase, that are abbreviated by H and N, respectively.) Resistance was particularly evident in Norway (see Box 11-1). Later in 2008, resistant virus reached South Africa: Of 23 samples for which neuraminidase activity was tested, all exhibited oseltamivir resistance. Another 45 South African isolates were tested for a viral nucleotide sequence change associated with resistance. All tested positive for resistance. Because none of the South African patients had been treated with drug, spread of resistant virus appeared to be occurring. By late 2008, resistance was prevalent in the United States.²³⁰

In 2008-09, the common subtypes of seasonal influenza were H1N1, H1N2, and H3N2. These subtypes usually caused mild disease; consequently, antibiotic resistance was an issue mainly for immunocompromised persons and the elderly. However, understanding the emergence and transmission of resistance is important for managing more aggressive forms of the virus. One of the key points is that oseltamivir resistance in the seasonal H1N1 virus strain did not interfere with transmissibility nor did resistance correlate with treatment: Resistant mutants moved easily from person to person. Because new strains of seasonal flu appear each year, oseltamivir resistance may disappear when the current strain is replaced. However, it could reappear with little warning.

Pandemic Influenza Can Be a Killer

A pandemic is an outbreak that spreads to multiple countries. The Spanish Flu pandemic of 1918–19 killed at least 40 million people worldwide within a 12-month period; more than 500,000 deaths occurred in the United States. An initial wave in the spring and early summer of 1918 was mild, whereas the second, in the fall of 1918, was severe. A third, less severe wave occurred in the winter and spring of 1919.²²⁹ Flu victims were not limited to the old and infirm: Members of the healthiest age group, 15- to 34-year-olds, died in staggering numbers. Indeed, killing young adults seemed to be characteristic of the 1918–19 flu.²³⁷ Smaller flu pandemics have also occurred. The 1957–58 Asian Flu killed nearly 1.5 million people worldwide, and the 1968–69 Hong Kong Flu had nearly 1 million victims. In 2009, a pandemic H1N1 strain originated from pigs in Mexico. To avoid embarrassment to Mexico, the pandemic was commonly called H1N1 Flu, which was confusing because much of the seasonal flu was caused by another H1N1 influenza virus. (The term Swine Flu lost favor because it led to erroneous notions about how influenza spreads.)

By mid-April, the 2009 pandemic had spread from Mexico to the United States and several other countries. Widespread panic did not develop. However, a lack of understanding surfaced in a variety of ways. Many people donned masks but failed to be properly fitted. That left large gaps between mask and face. Some countries banned pork imports even though the virus does not spread with food. Egypt experienced widespread slaughter of pigs for a virus that was spread directly from one human to another.

Experiences with influenza tell us what to expect from the next pandemic. Most important is the 6-month lag between virus emergence and vaccine availability. Second, obtaining and distributing sufficient vaccine may be difficult. We must not be lulled into complacency by the mildness of the 2009 pandemic. The panic associated with a highly lethal pandemic flu could overrun healthcare systems, cripple economies, and tear social fabrics as people fight for scarce medical resources. The Trust for America’s Health, a nonprofit, nonpartisan foundation, estimated in 2008 that a pandemic comparable to the one of 1918–19 would result in 90 million illnesses and more than 2.2 million deaths in the United States alone (www.healthyamericans.org). The number of deaths globally would be 50 to 80 million.²³⁸

Two other recent disease outbreaks sensitized governments to the potential devastation of an influenza pandemic. In October 2001, the United States experienced a deliberate release of the anthrax bacillus, and about 1 year later an epidemic of severe acute respiratory syndrome (SARS) jolted China and Canada (see Box 7-6). These two small disease events extracted such a huge social, economic, and political toll that health officials designated pandemic illness a threat to national security.

In the United States, local and federal plans were developed to minimize the impact of large disease outbreaks. Those plans included stockpiling antibiotics. Until late 2008, little attention was placed on how antibiotic resistance could magnify the problem, and in the 2009 N1H1 flu pandemic, governments initiated limited distribution of oseltamivir. Fortunately, resistance with H1N1 pandemic influenza appears to have remained low (as of December 2009, the prevalence was about 1%).

Avian Flu H5N1 Is a Candidate for Deadly Pandemic Flu

Since 2003, avian flu has received considerable attention as a potential pandemic virus. This version of influenza A, caused by subtype H5N1, is endemic to Southeast Asia. In the late 1990s, it began its worldwide spread via birds, and between 2003 and September 2008 it caused 387 documented human deaths.²³² The human death toll was small, largely because transmission of virus occurred from bird to human rather than from human to human. The unsettling number was the crude mortality rate—a staggering 63%.²³² Flu experts expect the mortality rate to drop considerably when the avian flu virus adapts to human-to-human transmission, but how much is unknown. In 2008, the H5N1 virus was still spreading effectively among birds, and infection of humans required exposure to high inocula.

As pointed out, influenza viruses evolve rapidly; strains that are seemingly fit can mysteriously disappear. In the case of H5N1, the virus has persisted in the bird population since 1997, a long time for an influenza strain.^239,240 During that time, many changes have occurred in virus recovered from poultry in Southeast Asia. For example, subtype H5N1 clade 2.3.4 became dominant in China in 2005 and then spread to northern Vietnam. (A clade is a group of viruses or organisms that descended from a single common ancestor.) In 2007, clade 2.3.4 replaced another Chinese clade that had successfully emerged in 2003.²⁴¹ As expected, the H5N1 virus is changing constantly, just like other versions of influenza virus.

Antibiotics May Play an Important Role in Pandemic Influenza

Strategies to counteract pandemic influenza depend largely on vaccine deployment, just as they do with seasonal flu. As of late 2009, no accepted vaccine for influenza H5N1 was widely available. (A vaccine had been made, but it was in short supply and was untested with human populations.) Vaccine makers were unsure which viral substrain to use for production. As pointed out, 6 months are required to prepare, test, and distribute a vaccine;²³⁸ consequently, our control over pandemic influenza has a serious time lag. Quarantine and isolation can help cover the vulnerable time, but disruption to society could be enormous if even a small fraction of people stayed home from their jobs. It is likely that anti-influenza drugs will be needed.

To implement an antiviral plan, the Strategic National Stockpile began accumulating millions of antiviral drug doses (www.bt.cdc.gov/Stockpile/). Millions more were to be acquired by businesses, hospitals, and other essential components of society. The plans call for a phased prioritization of antiviral prophylaxis, which means that the first persons treated would be those who render critical services. The same approach applies to vaccine deployment.

Both the adamantanes and the neuraminidase inhibitors limit infection by susceptible influenza viruses if used early in infection. Late in infection, when the body reacts to the virus with flu-like symptoms (respiratory discomfort, fever, aches, and pains), the viral load is already quite high. At that point, reducing it with drugs is presumed to have little impact. Indeed, with highly lethal viral strains, the uncontrollable health problems come from a massive inflammatory response by the human body. Consequently, waiting for categorical symptoms before administering drug is considered an ineffective strategy. Therefore, antiviral prophylaxis is a key public health strategy for a pandemic of lethal influenza.

Antibiotic Resistance Occurs with Avian Flu H5N1

Strategic use of antiviral agents requires that a pandemic virus be susceptible to existing antiviral drugs. Susceptibility may be true at the beginning of an outbreak, but it cannot be known in advance because the pandemic virus is unknown. The government plan also assumes that the virus will remain susceptible to the drugs during widespread use. Given what we know about influenza viruses, this is problematic. Adamantane resistance has been developing for several years (see Box 11-2), as these drugs have been used prophylactically with poultry in China.²³² By 2007, 30% of the avian flu isolates were resistant to the adamandanes,²³² making the agents unlikely to be useful in a pandemic. In 2009, the neuraminidase inhibitors were the only option if avian flu had started to spread among human populations.²³¹

Neuraminidase-mediated resistance tends to emerge from point mutations.²³² With oseltamivir and avian influenza virus, resistance following treatment has been attributed to conversion of histidine-274 to tyrosine. (This amino acid substitution fails to affect zanamivir treatment.²³¹) As the virus evolves in birds, antibiotic susceptibility changes in complex ways due to the evolution of the target protein and other interacting proteins. That complexity can cause virus isolates from different times and different geographic locations to differ in drug susceptibility, which makes antibiotic resistance in avian flu difficult to describe in a simple way.

If avian flu causes a human pandemic, the problem of antiviral resistance will be exacerbated by arming citizens with oseltamivir and asking them to prophylax: The public cannot be expected to always get the timing and dosing right. When resistant virus emerges from infected persons taking antibiotic, those strains are likely to spread, as we have seen with seasonal influenza virus. To use the agents effectively, we need rapid, accurate methods for determining virus susceptibility. Current methods are still cumbersome (see Box 11-3).

Bacterial Pneumonia May Create Another Resistance Problem

Influenza virus is only the beginning of the pandemic resistance problem. A major cause of flu-associated death is the bacterial pneumonia that follows flu.²³⁸ Indeed, in the 1918–19 pandemic, most deaths appear to have been due to follow-up bacterial pneumonia.^243,244 Although pneumonia-producing bacteria can be controlled with existing antibacterial agents, drug delivery systems will be challenged during a flu pandemic because we now rely on just-in-time supply chains. Developed countries have stockpiles of some antibacterials, but the optimal compound varies from one pathogen to another. For example, ciprofloxacin, which is quite effective with anthrax, is not recommended for pneumonia because its use with streptococcal pneumonia so often leads to resistance.⁷⁶ An alternative will be needed.

β-lactams, such as penicillin, and the newer fluoroquinolones are still widely used for infection by S. pneumoniae; consequently, they are likely to be administered prophylactically for pneumonia. That may create a problem with staphylococcal pneumonia, because MRSA is already penicillin resistant. Moreover, many HA-MRSA isolates are also resistant to fluoroquinolones. Thus, prophylaxis with β-lactams or fluoroquinolones will favor growth of MRSA. Vancomycin, one of the few effective antistaphylococcal drugs, requires intravenous injection, which cannot be easily administered in a pandemic setting. Moreover, vancomycin resistance occurs.⁶⁸ We conclude that antibiotic options may be quite limited.

Perspective

The striking emergence of antibiotic resistance in seasonal influenza virus is readily explained by the principles described in previous chapters. Factors likely to contribute are huge viral populations, agricultural and over-the-counter use of antibiotics, and dosing to cure rather than to prevent the emergence of resistance. What was responsible for oseltamivir resistance in the 2008 seasonal virus is unknown. In the case of neuraminidase inhibitors, exceptional care must be taken because the mutations causing oseltamivir resistance appear to move from person to person along with another change in viral RNA that improves transmissibility. Thus, continual selective pressure is not needed to disseminate resistant virus.

Some experts argue that pandemics are no more deadly than ordinary seasonal flu, as has been observed with the recent 2009 H1N1 pandemic.²⁴⁵ However, the human death rate is only part of the issue. The perceived rate, not the actual one, will drive infrastructure disruption. Although plans to deal with an influenza pandemic appear orderly and well conceived, they assume that citizens will act in an organized, cooperative manner. That was true in 2009. Unfortunately, human nature has a way of intervening in life-and-death matters. As a society, we are so accustomed to drug-based healthcare that a massive run on antiviral agents may occur during a pandemic with a deadly virus. For example, a sudden demand for ciprofloxacin occurred following the anthrax outbreak in 2001. Despite pleas from health officials, extensive hoarding of ciprofloxacin occurred, and the drug quickly disappeared from pharmacy shelves. Personal stockpiling of oseltamivir for influenza also occurs. In a survey comparing seasonal influenza cases and oseltamivir sales from 2002 to 2006, the two generally coincided in time. However, sales of oseltamivir rose uncharacteristically in the autumn of 2005, a time when both avian flu and oseltamivir received a burst of media attention.²⁴⁶ Another factor fueling uncertainty during a pandemic will be criminal elements. They sense easy profits, and counterfeit oseltamivir is produced for sale on the Internet. Low-quality antibiotics are a general problem for the emergence of resistance, because effective concentrations are low. (Totally inactive agents are the same as no treatment.) In the case of oseltamivir, a simple color test has been developed to determine whether a sample contains oseltamivir.²⁴⁷ We conclude that pandemic influenza associated with severe disease will create a complex, challenging healthcare problem. Nevertheless, individuals can take action to minimize problems. In the next chapter, we sketch some of those actions for a variety of infectious diseases and resistance issues.