Emerging Infections of International Public Health Importance
Kathleen Neuzil, MD, MPH
Dr. Kathleen Neuzil is a faculty member at the University of Washington, School of Medicine, Division of Allergy and Infectious Diseases, and does her clinical work at the VA Hospital. Her area of research interest is in vaccine preventable diseases, predominantly influenza, influenza epidemiology, and influenza public policy.
Influenza is a common acute febrile respiratory disease that affects all age groups and all segments of the population. People nonchalantly use the term “flu” to mean any number of diseases that cause nausea/vomiting, which may not be due to influenza. John Bartlett, a renowned infectious disease expert, said, “Influenza is probably the most underrated major pathogen in the developed world…one of the major infectious causes of death from a single microorganism.” In fact, influenza has the potential to cause enormous morbidity and mortality, and that is what makes it a real threat to public health:
Influenza has been described through the centuries, even though the influenza virus was not isolated until the 1930s. The hallmark of epidemic influenza is excess mortality, and it is a great disease to study epidemiologically. The study of influenza really started in 1837 with Robert Gray who was a physician in Dublin, Ireland, who noticed an association between excess mortality and the winter season. In winter there would be a rise in acute onsets of febrile respiratory disease, and more people died. Gray would go to the cemeteries in the spring and count the number of tombstones, and he first noticed this association that there were more deaths following winter seasons with epidemics of febrile respiratory disease. William Farr in London studied it more scientifically and coined the phrase “excess mortality”. It is not compensatory mortality; it is not sick or old people who are going to die anyway who were dying early. It is truly excess mortality that occurs due to influenza.
Selwyn Collins in the U.S. mapped out influenza epidemiology systematically over a 50-year period. Collins described that, although there would be some fluctuations, the standard baseline level for death was higher in winter than in summer. He recognized clinically and epidemiologically that more people are sick and more children are missing school in winter. When there were flu epidemics that ranged from 4-6 weeks, he noticed an association with excessive mortality.
There is nothing in the recorded history of man that has killed as many people in a comparable period of time than the 1918-1919 influenza pandemic. There was no other pathogen or natural catastrophe. In 1918, during a span of 6 months, the flu killed more people than the three years of WWI, and some believe that it helped accelerate the end of that war. Approximately 30-40 million deaths in the world were probably attributable to the pandemic. The 1918 influenza virus was isolated from two sources of preserved human tissue. During WW I, there were soldiers who died of a rapid hemorrhagic pneumonia, and some human tissue was preserved by the Armed Forces Institute of Pathology. In addition, the Alaska natives were decimated by this pandemic. Mortality rate was 70% in some of these villages because the natives had never seen influenza before. Bodies were buried in mass graves, and in permafrost the virus was preserved. Researchers from the Armed Forces Institute of Pathology were able to use viral RNA techniques to isolate H1N1 influenza virus type A from the site. Using both epidemiologic and molecular biology methods, researchers were thus able to show that both the WWI soldiers and the Alaska natives were affected by the same strain.
There was no vaccine back then, and I really appreciate the CDC’s emphasis on respiratory hygiene and wearing masks when you are sick. As we have seen, however, in 1918, wearing masks did not really prevent the spread of the pandemic. Here is a pessimistic quote from the American Journal of Public Health in 1918, where the editor described how to prepare for the pandemic: “hunt up your wood-workers and cabinet-makers and set them to making coffins. Then take your street laborers and set them to digging graves. If you do this you will not have your dead accumulating faster than you can dispose of them.”
It is interesting to look at the great pandemics of the last 100 years.
There was the 1892 pandemic in Massachusetts, the 1918 pandemic, and here are the pandemics of 1936 and 1957. An interesting feature of the 1918 curve is the W-shaped curve and the high death rate in young people. There are hypotheses about many young men gathering in close quarters during the war, but women actually had mortality that were as high (pregnant women had mortality rates up to 50%) and there has been no good explanation for this unusual curve.
Normally influenza epidemics have a U-shaped curve, with the highest mortality in the elderly and the very young. The mortality rates have decreased over time due to improved treatment of secondary diseases thanks to a better understanding of the disease pathogenesis. Influenza worsens underlying diseases and/or leads to bacterial pneumonia, but many of these diseases are treatable now with antibiotics and supportive care. In the U.S., there are still 10-40,000 deaths/year from flu, and 90% of them occur in people over the age of 70. So the total fatal cases and the rates are lower but the age patterns of the fatal cases remain unchanged; the highest mortality occurs in older people.
What is pandemic influenza?
A pandemic influenza is a global flu epidemic, and is caused by an emergence of an influenza A virus that is novel for the human population. Reservoirs for influenza A exist in many animals, such as birds, seals, and others. Comparatively, polio control is easier, because polio is an exclusively human disease. If there are human populations that are in contact with animals, viruses can re-assort to produce novel progeny with the ability to spread within the human population, there is a potential for a pandemic.
The World Health Organization (WHO) has a pandemic alert system. Through the global influenza surveillance they look for:
To prevent a potential pandemic, you need surveillance centers globally. Even 10 years ago we had surveillance in the U.S. but not in Asia, where many influenza strains originate. Labs are also needed to map the genes to see if it is an influenza A virus with a novel surface protein. Strong communication and vigilance are equally important, and WHO has an email influenza alert system. As it was done for SARS, in addition to isolation and containment, intervention is necessary. With a novel antigen, there needs to be rapid vaccine production or distribution of influenza antiviral medications. Every year, it takes ~ 6 months for vaccine development (via embryonated hens’ eggs) after surveillance designates which strains are likely to circulate. It then takes some time to distribute the vaccine. In a pandemic situation there may not be enough time to develop and deliver a vaccine in time before the virus has circulated the globe. Thus we need rigorous surveillance, rapid communication, emergency preparedness (with an isolation and quarantine plan), and a plan for intervention with vaccinations/antiviral medication.
The Influenza Virus
This is the influenza virus (Figure 8: Schematic of Surface Antigens of Influenza A). The surface proteins—the hemagglutinin and neuraminidase antigens—predict the virulence. Hemagglutinin is a center for attachment of the virus to human epithelial cells; if you block hemagglutinin with antibody or with vaccines you can prevent infection. Neuraminidase is important for the virus to spread from cell to cell; by blocking neuraminidase you cannot prevent infection but you can limit spread of the infection. Because of the segmented RNA genome, you can have different flu viruses mixing together in certain animal reservoirs. If you have a human flu virus and a bird flu virus in the same animal, such as swine, and if you get the right mix that confers a new hemagglutinin but retains the ability to spread in humans, then there is concern.
After 1918, the next great pandemic was in 1957. The strain circulating before 1957 was H1N1, but in 1957 H2N2 began circulating. Since the H2N2 strain was new to the population there was a selective advantage for this strain to spread and infect people. In 1968 the H3N2 strain appeared, a combination of avian and human antigens. In 1977 the H1N1 strain reemerged--identical to human epidemic strains from the 1950’s. Since 1977 we have had both H1N1 and H3N2 as well as influenza B strains circulating in the population. Influenza B viruses can cause epidemics but do not undergo the same degree of mutations and therefore have not been associated with pandemics.
In 1997 in Hong Kong there were 18 confirmed cases and 6 deaths from an influenza due to a novel strain, H5N1, originating in birds. Many chickens fell ill with flu and there was a mass slaughter of poultry in order to control the outbreak. While this may seem like a drastic public health approach, there were no further humans cases in the ensuing months. Since 1997, there have been additional mass cullings of chickens for strains H5N1 and H9N2 in Hong Kong and currently in Southeast Asia for H5N1 (with victims from Vietnam and Thailand).
During the last 20 years, prior to the outbreak in Hong Kong, there have been isolated cases of people getting infected with animal flu viruses, predominantly in the U.S. or in the Netherlands.
These small isolated events are likely occurring elsewhere, but they are being recognized more in the U.S. and the Netherlands due to their extensive surveillance. In Hong Kong, the 1997 outbreak was difficult not to recognize as it decimated the poultry industry.
What we are seeing now is partly due to increased awareness. Last year there were some isolated events in Hong Kong and the Netherlands. In such situations, there is concern regarding the ability of influenza to spread person-to-person. It is important to know if infected individuals had direct contact with an animal of concern, which would indicate an isolated animal-to-human event. The investigators went out in a “circle” around the infected people to see if family members and health care workers who did not have direct contact with animals were infected or not. Some health care workers were in fact found to have antibodies, which indicate that there was some limited human-to-human transmission in the Hong Kong and Netherlands outbreaks.
It is not entirely clear why many influenza outbreaks appear to originate in Southeast Asia. There are other places with high animal-to-human contact, but in Southeast Asia you have high human population density together with high animal population density. In Guangdong province, the people are poor and their animals are in close proximity to humans. There are also many intermediate animals, such as swine, that act as a “mixing vessel” for different flu viruses. The region is also subtropical in climate, and data from Hong Kong and southern China suggest they have influenza circulation year-round, thereby increasing the chances of exposure for both humans and animals. Also, live animal markets are present in these regions. Live poultry are present not only in rural areas but also in dense urban markets.
In the U.S. and most temperate climates, influenza is seasonal with distinct peaks during the winter. There are other viruses in temperate climates, such as respiratory syncytial virus (RSV) and parainfluenza virus, which circulate also in winter.
This is why it can be clinically difficult to recognize flu.
An Influenza Epidemic in Houston
Paul Gleason is an influenza epidemiologist and mapped out the time course of community illness in Houston, Texas, USA. He showed the first effect from influenza is student absenteeism in schools.
Emergency room visits increase for respiratory diseases and pneumonia. Children then spread flu to their parents and others, leading to work place absenteeism and an increase in adult hospital admissions. Death from influenza typically lags by ~2 weeks after the initial signs of outbreak hits an area. If you do passive surveillance in schools, a 10% school absenteeism range is a good marker of influenza in the community. Public health departments thus ask school nurses regarding absenteeism, and sometimes that is the first way they pick up influenza.
Influenza Attack Rates in Children
Children in general have higher influenza attack rates (usually 20 - 30%) than adults. This was a 25-year study and we prospectively followed up children aged <5 years who were enrolled in a clinic at some point from 1974 through 1999.
Viral cultures were obtained when the children presented with clinical illness to determine the incidence of laboratory-confirmed influenza illness. The attack rates are variable yearly as are the viral strains predominant in the community. In 1978, it was only 1% of children who were sick enough with flu to come see the doctor, while in 1993 it was 19% of children. If there is low influenza activity for a few years, a higher proportion of people in the population become susceptible, and this is often followed by a severe influenza season with high attack rates. However, it is still difficult to predict influenza patterns because you can have a severe influenza type B year followed by a severe type A year.
This is another study we did over a 20 year period.
If you assume a 30% attack rate of influenza in children, then only a percentage of these children, or 6-14% of all children, will be ill enough to warrant an outpatient visit. Also, it is hard to tell whether younger children seem to be sicker or their parents are more concerned to bring them to the doctor. This Figure also shows that at least 50% or more of these children will receive a course of antibiotics. The overuse or misuse of antibiotics (influenza is a viral infection) is a concern. It is unclear whether the concerned pediatrician is unsure of the illness and hands the worried parent an antibiotic or the flu really is leading to acute, secondary bacterial infections.
Hospitalizations Attributable to Influenza
These are some data over about a 20 year period from Tennessee Medicaid and looks at how many hospitalizations are attributed to flu.
This information helps determine who should receive the influenza vaccination. Note how similar this is to those pandemic influenza curves for mortality. This is hospitalization, but you see the same U-shaped curve where you have the highest rates in the very young and the very old. High risk people, such as persons with emphysema, asthma, cancer, babies born prematurely or with congenital malformations, etc., get much sicker with flu than low risk, healthy people. Most deaths, however, still occur in the elderly.
Influenza Vaccine Policy
In this country, we put a lot of emphasis on old people getting flu vaccine, which I strongly believe in. But we should also be giving influenza vaccine to the very young. In fact, starting next year, there is a recommendation that children 6-23 months of age receive the influenza vaccine (high risk people no matter what age should continue to get vaccinated). The vaccine is not licensed for children younger than 6 months of age because it has not been adequately tested.
The CDC reports influenza activity through weekly MMWR reports of pneumonia and influenza deaths using an epidemic threshold. Influenza seasons are designated as severe, moderate, or mild depending on the number of weeks the threshold is exceeded and by what amount. By emphasizing mortality, the elderly are targeted and protected, but looking at solely mortality does not address all of influenza’s impact.
If you look at any vaccine program, there are a number of parameters considered:
Very rarely will we give a vaccine that is unsafe. A good example is smallpox—although there are some adverse reactions to the vaccine, the mortality rate is 30%, so you tolerate the adverse effects. For the most part, vaccine safety is emphasized with supportive evidence of clinical effectiveness. And, we also acknowledge that the flu vaccine can reduce the large burden of influenza on society. On the contrary, we are not going to give vaccine for anthrax or smallpox to an entire population, because there is not enough exposure for the average person to justify a vaccine.
Cost-effectiveness is another issue. For example, cost studies showed that it is cheaper for the government to pay for the chicken pox vaccine and prevent it rather than not have a vaccination program. Another concern is the epidemiology of transmission: are we vaccinating certain groups to protect other high risk groups? Should we vaccinate children to protect their grandparents against influenza?
Vaccine production can be timely and difficult. There is lag time in surveillance, and a new vaccine must be produced every year. In the U.S., we have to start making the vaccine in March. When a new strain arrives, the old strain is still around, but there are more susceptible people to the new strain, so it becomes more predominant. So you have several strains circulating. This year, we had some strains circulating that were in the vaccine, but the predominant strain was not included. So we need new, more rapid influenza production techniques, or we need a vaccine that can protect against a broader range of strains. For instance, the new intranasal vaccine, a live vaccine, may theoretically induce broader immunity and protection.
The WHO board decides the flu vaccines twice a year, for the northern and southern hemispheres. The U.S. FDA generally follows the WHO recommendation. Sometimes they may use a slightly different strain, but generally the vaccines used throughout the world are the same.
The main focus on influenza is prevention and control, and our policy is to keep
people from getting complications of flu.
Persons at high risk for complications that should be immunized are:
The second emphasis that we have concerns people who can transmit to those at
high risk. Thus the following persons should also be immunized:
Vaccination of health care workers
Potter et al conducted a study where they looked at influenza vaccination of health care workers and patient mortality in long-term-care hospitals. The study was conducted on 1059 residents in 12 geriatric long-term-care hospitals in Glasgow, Scotland. The flu vaccination rates were quite low for these patients. They randomized hospitals for health care workers to get vaccine or no vaccine and then looked at patient mortality. Potter et al wanted to know if vaccinating health care workers would prevent spread to the elderly patients and reduce their mortality.
Over the four months period of winter, there was an overall 16% patient mortality rate. The mortality rate was the same regardless of the patients’ vaccination status.
When assessed by hospital, Potter et al found that in the hospital where staff were vaccinated, the mortality rate was significantly reduced among the patients.
The residents are not going out to the community, and the only way they are going to be exposed to flu is if somebody brings it in, such as family members or health care workers. Whether this plan works in a community where we have multiple exposures at all times is hard to say.
Other groups to consider for vaccination are:
The CDC recently changed the high risk category to people age 50 and over. The reason for age 50 was the recognition of a linear increase in risk and prevalence of diseases that place persons above age 50 at increased risk. However, when there are vaccine supply issues, the CDC prefers to vaccinate a 65 year old over a 50 year old. There are some data from studies that show that maternal antibody either via vaccine or breastfeeding may protect the infant, but the data are not very reliable. We do recommend that pregnant women get flu vaccine both to protect themselves and hopefully provide some protection to the infant.
People who should not be immunized are persons who have anaphylactic hypersensitivity to eggs or other components of the vaccine and persons with acute febrile illnesses.
Overall, influenza vaccine efficacy is lower than measles, varicella or tetanus. There are mismatched years in which the strains in the vaccine do not match the strains circulating in the environment. Thus the efficacy can vary annually. Age group is also a factor. In healthy children and adults, the vaccine efficacy is 70-90% in preventing culture-confirmed influenza. In the elderly, the efficacy drops to 30-70% in preventing hospitalization for pneumonia and influenza, and for the frail elderly it drops to 30-40% in preventing illness (50-60% in preventing hospitalization and 80% in preventing death). When you get older, your immune system weakens, but the vaccination still helps in preventing serious complications and attenuates illness. Moreover, considering the burden of flu, and the relatively inexpensive cost of the vaccine, it is better to prevent even 30%.
The problem with flu is that it is not obvious like chicken pox. Parents criticize the flu vaccine because vaccinated children still miss school with a cough/cold two months later. However, the illness could have been caused by another virus, such as RSV. This causes a lot of public health confidence problems. If you give your child a measles vaccine, and everybody else in school that did not get vaccinated gets a rash that is positive feedback for the vaccine. We do not have this positive feedback with influenza. Measles is a great vaccine success story in the U.S., but it is not a prototype for an influenza vaccine campaign. You can give a measles vaccine any time of the year and you do not have to give it annually. Also, flu is not an exclusively human disease like measles, and thus cannot be eradicated.
If we look at our flu vaccine coverage rate with public health campaigns, we do decently in older white people, but we only vaccinate 20% of high risk children.
This Figure shows our goals for coverage among the high risk populations. We give about 80 million doses of influenza vaccine which is one of the highest flu vaccine rates in the world but still less than one-third of the US population. Another challenge is the appropriate window of time for vaccination, from September to November. In 1999, we had 44 million doses in September, and by December we had 76 million doses, but if flu season comes early, it can be too late.
Influenza Vaccine and Guillain Barre Syndrome
In 1976 there was a swine flu campaign, and it became associated with Guillain Barre Syndrome. The annual incidence of Guillain Barre is about 10-20 patients per 1 million adults. In 1976, there was swine influenza in a soldier, and the US Public Health Service and the CDC launched a mass immunization campaign. However, there was no convincing evidence of person-to-person spread. There was a higher than expected incidence of Guillain Barre Syndrome in 1976 related to the swine flu vaccination, and in other influenza seasons, there was an excess of ~one additional case per million persons vaccinated (1992-93 and 1993-94). This was quite unfortunate for what it did for the confidence of flu vaccine. There may have been some reporting bias here, if you got Guillain Barre few weeks after you got the flu vaccine. We also know now that the most common cause of Guillain Barre is due to a diarrheal infection, Campylobacter jejuni. It is unknown how many cases may have had a diarrheal infection when they were vaccinated.
Vaccination of Children
Children are important in the spread of flu, and a study looked at whether immunizing day care children could reduce flu in the home. A randomized control trial by Hurwitz et al involving 127 children (aged 24-60 months) and their 328 household contacts among 10 day care centers was conducted 1996-1997. Sixty children received influenza vaccine while 67 received hepatitis A vaccine. The researchers showed that looking at any respiratory illness as an outcome, the vaccine efficacy was ~16%. But this could include RSV, parainfluenza virus, and other agents. When they made their criteria more sensitive and looked at fever, a hallmark of influenza, and then specifically high fever, they saw that you could prevent 42% and 47%, respectively, of cases of flu in adults in the home by vaccinating the children.
Monto et al conducted a study in 1970 to see if vaccination of schoolchildren could modify a flu outbreak in the community. They took a small community in Michigan and immunized 90% of the schoolchildren in the community and compared all respiratory illnesses to a control community with similar demographics without any vaccination.
There was a reduction of overall respiratory illness (includes all respiratory illness, only a portion of this is flu) in every age group where the children were vaccinated--not only in the children that were vaccinated, but also in younger children and adults. Thus there is some suggestion that if you target vaccines at children, you may be able to make a difference at the greater community level.
In fact, in Japan, Reichart et al showed that high flu vaccination rates in school children were associated with a decrease in mortality in the elderly. From 1977 to 1987, flu vaccine coverage rates among Japanese schoolchildren were >80%; rates began to drop in late 1980’s and early 1990’s. In this ecologic study, the number of monthly deaths in Japan was obtained for 1949-1999. The winter excess mortality was sharply reduced from the mid-1970s to 1992, when they had the schoolchildren vaccination program. The excess mortality resumed when Japan terminated the vaccination program. In Japan, there are many more multigenerational families, so whether this strategy would work in the U.S. is difficult to know, but vaccination of 80% of school children appeared to have averted ~30,000 deaths per year in the elderly in Japan.
Early in an epidemic, if you have a limited amount of vaccine, our country’s current policy is to vaccinate the old and sick. An alternative argument is to vaccinate the people moving about and spreading the disease if you have a limited vaccine supply. In addition, vaccines are less effective in elderly or debilitated populations. An option for the elderly and others at high risk is antiviral medication. Vaccination is the best way to control influenza, but if we have a pandemic, or we have an outbreak in a nursing home and we cannot make the vaccine fast enough, antivirals are the best alternative. With prophylaxis, antivirals are 70-90% effective and they work against all strains that have been tested. However, prophylaxis must be taken on a daily basis to be effective, so adherence is an issue as compared to a “one-shot” vaccine.
There are limitations of antiviral medications for treatment of influenza:
Amantadine and Rimantadine antivirals have the following features:
Neuraminidase inhibitors (Oseltamivir and Zanamivir) have the following features:
Influenza is an emerging infection that is a serious public health concern. It
is a major cause of morbidity, mortality, and lost productivity worldwide. Flu
vaccines should remain the cornerstone of flu control efforts, but there are
many problems with the current vaccine programs. Antivirals are an important
adjunct to influenza control, particularly during vaccine mismatch years.
Finally, influenza preparedness plans should be multifaceted and include
surveillance, prevention and communication components.