Announcement

Collapse
No announcement yet.

Influenza in the acute hospital setting

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Influenza in the acute hospital setting

    The Lancet Infectious Diseases 2002; 2:145-155
    DOI:10.1016/S1473-3099(02)00221-9
    Influenza in the acute hospital setting

    Cassandra D Salgado a, Barry M Farr a, Keri K Hall a and Professor Frederick G Hayden a

    See Errata

    Summary
    Epidemiology of nosocomial influenza in the acute care setting
    Nosocomial impact
    Transmission
    Nosocomial aspects
    Duration of contagiousness
    Detection and diagnosis
    Prevention and control
    Immunisation
    Antiviral chemoprophylaxis
    Oseltamivir and zanamivir
    Search strategy and selection criteria
    References

    Summary

    Influenza poses special hazards inside healthcare facilities and can cause explosive outbreaks of illness. Healthcare workers are at risk of acquiring influenza and thus serve as an important reservoir for patients under their care. Annual influenza immunisation of high-risk persons and their contacts, including healthcare workers, is the primary means of preventing nosocomial influenza. Despite influenza vaccine effectiveness, it is substantially underused by healthcare providers. Influenza can be diagnosed by culturing the virus from respiratory secretions and by rapid antigen detection kits; recognition of a nosocomial outbreak is important in order to employ infection-control efforts. Optimal control of influenza in the acute-care setting should focus upon reducing potential influenza reservoirs in the hospital, including: isolating patients with suspected or documented influenza, sending home healthcare providers or staff who exhibit typical symptoms of influenza, and discouraging persons with febrile respiratory illness from visiting the hospital during a known influenza outbreak in the community. (Note: influenza and other respiratory viruses can cause non-febrile illness but are still transmissible.) The antiviral M2 protein inhibitors (amantadine, rimantadine) and neuraminidase inhibitors (zanamivir, oseltamivir) have proven efficacy in treating and preventing influenza illness; however, their role in the prevention and control of influenza in the acute hospital setting remains to be more fully studied.
    Back to top

    An important cause of morbidity and often mortality in most communities every winter, influenza poses special hazards inside healthcare facilities. Because of its short incubation period and efficient respiratory spread from person to person, it can cause explosive outbreaks of febrile respiratory illness. The hospital patient population has serious underlying illnesses, making influenza more lethal in this setting.1?5
    Healthcare workers have increased risk of acquiring influenza during known outbreaks since they are exposed to infected individuals in the community as well as hospitalised patients with influenza. Whether infected at work or in the community, healthcare workers may become an important source of influenza for their patients.4,6,7 For example, a serological study in the UK, where healthcare worker immunisation had not been the standard of care, found that up to 23% of healthcare workers contracted clinical or sub-clinical influenza during community outbreaks.8 Up to three-quarters of healthcare workers with influenza-like illness have continued to work while ill9 and thus may perpetuate transmission and foster outbreaks. Illness in healthcare workers may also disrupt patient care and result in higher hospital expenditures.
    Optimal prevention of nosocomial influenza should involve multiple strategies that focus upon reducing potential influenza reservoirs in the hospital. For example, campaigns encouraging healthcare workers to get vaccinated is one way forward (figure 1).We will review relevant aspects of influenza epidemiology in the acute-care setting, its impact, its modes of transmission and reservoirs of infectivity, and methods of nosocomial influenza detection, control, and prevention.


    Click to enlarge image


    Figure 1. Poster displayed at University of Virginia showing proportion of UVA healthcare workers receiving influenza vaccine.


    Epidemiology of nosocomial influenza in the acute care setting

    Hospital admissions

    The annual number of influenza cases and subsequent influenza-associated hospital admissions, complications, and deaths in the community are dependent upon the pattern of disease and circulating viral strains. Annual influenza epidemics typically affect 5?10% of the adults in the community and result in hospitalisation rates of 2?4/10 000 in otherwise young healthy adults.10,11 However, hospitalisation rates are much higher among infants and young children, the elderly and certain high-risk populations.12?14 In influenza epidemics from 1968?1997 in the USA, the number of disease-related hospital admissions ranged from 16 000?220 000 per epidemic with over half of these hospitalisations occurring in those <65 years of age.12?15 Many of those hospitalised are considered to have exacerbations of underlying disease and are not always recognised as having intercurrent influenza infection, so they may serve as potential sources of spread.
    Influenza-associated deaths are often the result of pneumonia or exacerbations of cardiopulmonary conditions, and the elderly population currently accounts for approximately 90% of all influenza-related deaths.16 Excess influenza-related deaths occurred in 19 of 23 influenza epidemics with rates ranging from 30?150/100 000 persons in those older than 65 years in the USA.15 Of note, epidemics of influenza A H3N2 subtype viruses have been associated with greater hospitalisation and mortality rates. In addition to causing substantial morbidity and mortality, epidemics often have a deleterious, sometimes overwhelming, effect on healthcare delivery and cause marked increases in healthcare costs. This results not only from a sudden surge in the number of hospital admissions over a relatively short period of time, but also disruption of services due to an increased number of ill caregivers.15,17 Medical cost-containment efforts, resulting in decreased numbers of available beds and staff, could exacerbate this problem and lead to shortages of acute care hospital beds during epidemics.
    An even more dramatic and devastating effect on the healthcare system has been described with pandemic influenza. The pandemic of 1918?1919 was responsible for nearly 40 million deaths and over 500 million infections worldwide.18 It disrupted essential community services including healthcare delivery in many communities.19 The world-wide cumulative number of deaths in the 1957 Asian and 1968 Hong Kong pandemics was estimated at nearly 6 million.18 Furthermore, pandemics are associated with excess hospitalisations and mortality in adults below the age of 65 years.16 One of the first well-documented descriptions of nosocomial influenza involved a severe outbreak on a general medical service during the 1957 pandemic.20 Estimates indicate that the next pandemic will cause between 89 000 and 207 000 deaths and between 314 000 and 734 000 hospitalisations in the USA alone.21
    Back to top

    Nosocomial impact

    During periods of increased influenza activity in the community and subsequent higher numbers of patients with influenza in hospitals, the risk for nosocomial transmission from infected patients increases. In addition, nosocomial influenza may be acquired from healthcare workers, other hospital employees, and visitors. Influenza can cause sporadic or epidemic disease in hospitalised patients, and nosocomial outbreaks have been described in regular hospital wards, neonatal and adult intensive care units, transplant centres, and in nursing homes and other chronic care facilities.
    Nosocomial influenza outbreaks are not uncommon but are likely substantially under-reported. Outbreaks usually occur during periods of community influenza activity and are associated with notable morbidity, increased cost of care, and excess mortality.1?5,17 In a review of 12 nosocomial influenza outbreaks2?4,6,7,22?29(table), the outbreak duration ranged from 2?69 days (median 7 days) and reported patient attack rates ranged from 3?50% on epidemic wards and 0?7?20% throughout the hospital. Healthcare worker attack rates were infrequently reported but ranged from 11?59% among those caring for patients with influenza. Much lower healthcare worker infection rates have been documented at our hospital (less than 2% in any season during the past decade) where workers have been encouraged to be vaccinated, to report febrile respiratory illness, to be screened for influenza, and sent home when ill to prevent spread of infection to patients.30 Most nosocomial outbreaks are caused by influenza A with nosocomial influenza B infections occurring primarily in hospitalised children; however, influenza B adult outbreaks have been reported and one outbreak affected 20% of hospitalised adults, whose average age was over 60 years.4 Although valuable information is gained from review of these outbreak investigations, most were not prospective studies and may not be representative of all hospital experience with influenza.

    Click to view table




    Table 1. Description of reported nosocomial outbreaks in the acute-care setting


    Mortality rates during nosocomial outbreaks vary according to the affected patient population and circulating strain. In neonatal ICU outbreaks, mortality may be low, except for deaths occurring in babies with severe underlying cardiopulmonary disease.24 In acute-care facilities, and geriatric hospitals, median mortality has been 16%,2,4,6 whereas in specialised populations such as transplant or ICU patients, the mortality can be 33?60%.3,29 The immunocompromised population including transplant recipients, those receiving chemotherapy for malignancies, and those with advanced HIV disease deserve special consideration as illness secondary to influenza is often severe. Up to 70% of influenza infections in immunocompromised transplant and acute leukaemia patients have been nosocomial in origin in several studies.31?33 Pneumonia and death are common complications of influenza in immunosuppressed children, bone-marrow-transplant recipients, and those receiving chemotherapy for acute leukaemia.33?37 Patients at greatest risk of these complications seem to be those with the highest degree of immunosuppression.36
    Information regarding the economic impact of nosocomial influenza is limited, but the financial burden can be substantial, in part because of prolongation of hospital stay and additional diagnostic and therapeutic interventions for affected patients. Excess hospital costs are also due in part to absenteeism of healthcare workers who become ill during an influenza outbreak, since hospitals must cover the salaries of the ill workers in addition to their replacements.17 One study estimated a mean excess hospital cost of over $7500 per episode of nosocomially acquired influenza.17
    Back to top
    Transmission

    General principles

    The influenza virus spreads from person to person by respiratory droplets. When a person coughs or sneezes, particles of varying sizes are created ranging from large particle droplets that are filtered by the nose and pharynx to small particle droplets that reach the alveolus. The observation that influenza can be induced by intranasal drops38 suggests a role for direct contact spread, but several lines of evidence indicate more efficient transmission via smaller particle aerosols. In experimentally infected volunteers, the human infectious dose is at least 10?100 fold lower with inhalation of small particle aerosols delivered to the lower respiratory tract than with direct instillation of nasal drops.39 Further, intranasally administered influenza virus uncommonly causes cough or lower respiratory tract manifestations in experimentally infected volunteers,40 whereas early onset and protracted cough indicative of tracheobronchitis is characteristic of natural influenza.41
    Studies of differing routes of administering antiviral drugs also suggest that intranasal inoculation is not the usual route of natural transmission of influenza. Intranasal zanamivir and intranasal interferon are protective against intranasally introduced influenza under experimental conditions, but both fail to prevent natural influenza in ambulatory persons.42?44 By contrast, inhaled zanamivir is protective against natural influenza in household and nursing home contacts.45 Because of the range in natural respiratory droplet particle size and a resulting gradient in where particles land in the respiratory tract between the nasopharynx and the alveolus, it is possible, however, that droplets of intermediate size (ie, those capable of passing the nose but not able to reach those more than about 5 feet from the coughing individual) could have an important role in influenza transmission. Of note, influenza viruses retain infectiousness after aerosolisation for at least 24 h at room temperature under conditions of low relative humidity.46
    Some epidemiological observations also point to transmission through the air. One point source outbreak due to an infected passenger on a delayed aeroplane with a non-functioning ventilation system found a 72% clinical attack rate for people throughout the passenger cabin over the following 4 days; the risk of illness was dependent upon the amount of time spent on board during the repairs of the system.47 This outbreak is often used as evidence of smaller particle droplet transmission, although it remains possible that large particle droplet spread also had a role in transmission during this outbreak as passengers moved freely about the passenger compartment.47,48 Such observations suggest that most naturally occurring influenza infections occur by inhalation of droplets or droplet nuclei rather than direct intranasal inoculation (eg, from contaminated fingers).
    Back to top

    Nosocomial aspects

    In general, hospitalised patients with influenza are maintained in private rooms but are not placed into negative pressure isolation. Despite this, only rare instances of spread outside the isolation room have been recognised. For example, we have not recognised time-space clusters of cases of nosocomial influenza during the past 15 influenza seasons at the University of Virginia despite putting most cases of influenza into private, positive pressure ventilation rooms. However, when susceptible patients or groups are cohorted with a infected index case, spread through the air seems to be responsible for some transmission. One study assessed influenza virus cross-infection during an influenza season in paediatric wards and found that patients admitted to wards with a greater proportion of beds in an open area were more likely to acquire nosocomial influenza than those who were admitted to wards with a greater proportion of private beds.49 While this suggested spread through the air, the data cannot clearly distinguish between large and small particle droplet spread because of lack of information about the spacing of the beds and any time-space clustering of the illnesses. Furthermore, one of the four wards in the group with more nosocomial infections had 100% private rooms, which accounted for 29% of all of the infections in this group.49
    The rapid onset and explosive nature of nosocomial influenza outbreaks with secondary cases often occurring within hours of each other have been used to suggest that smaller particle aerosol transmission must be involved.7,24 In one of these outbreaks, two culture-proven and one suspected nosocomial case occurred within hours of one another, but each patient had been cared for by a healthcare worker with influenza.7 In the other outbreak four infants developed illness over a 4-day period; two of the babies' cribs were adjacent to one another suggesting possible transmission. However, four neonatal ICU nurses were sent home just before onset of the cases for flu-like illness, although they had reportedly not worked with the affected infants.24 While the sources and transmission modes of influenza are unclear in such reports, they implicate healthcare workers.7,24 Influenza viruses can retain infectiousness on non-porous surfaces for 24?48 h and for at least 5 min on hands,50 and one nursing home outbreak implicated transmission by the hands of healthcare providers.51 It seems reasonable to assume that multiple routes of nosocomial transmission are possible.
    Back to top
    Duration of contagiousness

    An important issue in acute-care settings is the type and duration of isolation for patients with documented or suspected influenza. The principal reservoir of human influenza is the respiratory tract, particularly nasopharyngeal secretions and sputum of infected people. The period of communicability is estimated by the duration of recovery of infectious virus from respiratory tract samples but is likely to be heavily influenced by the presence of symptoms that generate infectious aerosols (eg, coughing). Viral shedding may begin at least 1 day before clinical illness and the duration of shedding depends on the age and immune status of the affected person. Outpatient children commonly shed for 7 days or longer, whereas non-hospitalised immunocompetent adults may shed for 4?5 days after symptom onset.52 Institutionalised elderly adults have shown more prolonged virus recovery (for up to 7 days),53 and nosocomially infected infants shed virus for up to 21 days.23
    Immunosuppressed patients often also shed for prolonged periods of time. For example, one 7-year-old girl with advanced HIV disease had nasopharyngeal shedding of influenza for at least 9 weeks54 and among influenza-infected adult bone-marrow-transplant or leukaemia patients, the median duration of shedding was 7 days but ranged up to 44 days.55,56 Of note, among rimantadine recipients, 83% of those shedding virus for 3 days or more had resistant variants. The duration of virus recovery is not well characterised in hospitalised adults but many have detectable virus in upper-respiratory-tract samples at 5 days after admission.57 The duration of viral replication in the lower respiratory tract has not been studied rigorously, nor has the effect of antiviral therapy on viral recovery been critically assessed in hospitalised patients. Prolonged shedding has important implications for nosocomial spread, in that such patients may serve as reservoirs of influenza virus, including drug-resistant variants. Sequential virological monitoring of patients at high risk for protracted viral excretion may be necessary.
    Back to top
    Detection and diagnosis

    Recognition of an outbreak is essential in order to employ control efforts such as isolation of patients suspected of having influenza, discouraging visits by those with suspected influenza, and sending home ill healthcare workers. The development of unexplained fever and/or respiratory illness in a hospitalised patient should raise the possibility of nosocomial influenza during influenza season. Hospitalised elderly adults may exhibit atypical features including gastrointestinal manifestations and altered cognition. In infants, a sepsis syndrome may occur with apnoea, lethargy, and poor feeding.28 Influenza may occur sometimes outside the usual seasonal patterns, so outbreaks of unexplained febrile illness during summer months warrant appropriate diagnostic studies.
    Influenza can be diagnosed by culturing the virus from respiratory secretions (nasal swabs, throat swabs, or nasal washes), or by more rapid influenza diagnostic tests for antigen (immunoassay or immunofluoresence), neuraminidase enzymatic activity, or viral RNA by reverse transcription PCR. The performance of one enzyme immunoassay, Directigen FLU-A (Becton Dickinson Microbiology Systems, Cockeysville, MD) and a direct immunofluorescence test, IFA, were compared with culture isolation of influenza A in geriatric patients.58 Directigen FLU-A had a sensitivity of 86?8% and a specificity of 99?1%, and IFA had a sensitivity of 92?5% and a specificity of 97?2%.58 By providing same-day diagnosis, rapid testing allows for early detection of nosocomial influenza outbreaks and enables infection control departments to institute appropriate containment measures throughout the hospital to prevent further spread. However, sending specimens for viral culture is important because only culture isolates can provide information regarding influenza subtypes and strains and potential drug resistance. Furthermore, the available rapid diagnostic tests vary in their performance characteristics according to assay type, sample type, time of sample collection, and age of the patient. When compared with influenza isolation, immunofluorescence assays are 50?90% sensitive for viral detection58?61 and enzyme immunoassays are 50?80% sensitive.62?65 In general, the sensitivity of rapid diagnostic tests is higher in children than adults and higher for nasal samples than throat specimens.59 False-positive results occur with some commercial antigen assays and have resulted in incorrect diagnosis of influenza as a cause of several outbreaks of respiratory illness. The positive predictive value of these tests outside known influenza epidemics may be low.60?63
    Annual prospective surveillance of influenza activity in the community helps to define the period of risk for hospitalised patients. Passive surveillance by employee health services may not be helpful as the only method of community influenza surveillance because it often underestimates the amount of active influenza illness in the community. Once influenza is known to be circulating in the community, we recommend routine influenza testing in patients presenting to hospitals for admission with an acute febrile respiratory illness. Hospitalised patients who develop symptoms compatible with influenza during a known outbreak in the community should also be tested. Pseudo-outbreaks due to laboratory contamination should be considered when unusual isolates or recovery from persons with atypical illness occur.66
    Back to top
    Prevention and control

    Interrupting transmission of virus

    Effective control of outbreaks in acute-care facilities is challenging since several reservoirs for transmission exist, including patients, providers, staff, and visitors. Once a community outbreak has been recognised, patients presenting for hospital admission with febrile respiratory illness should be tested for influenza and placed into droplet precautions (private room, mask for healthcare workers and for patients when they leave the room) until their influenza status is known (figure 2). Similarly, droplet precautions should be employed for hospitalised patients with documented or suspected influenza.15 As influenza rapid antigen tests are relatively insensitive, patients with suspected influenza and a negative rapid antigen test should remain in droplet precautions for at least 5 days pending results of influenza cultures before removal from isolation. When appropriate, it is also important to consider and screen for alternative diagnoses, such as respiratory synctial virus, adenovirus, Mycoplasma pneumoniae, and others. Rapid tests may be useful in situations where patient cohorting becomes necessary. Patients with documented influenza of the same type may be placed in the same room or cohort ward. To prevent further spread, visitors with fever and respiratory symptoms should be discouraged from entering the hospital. This can be accomplished by public service announcements in local newspapers, on television news, over radio broadcasts, and on memos posted around hospital entrances. Another nosocomial reservoir is healthcare providers and other hospital employees who have patient contact. Sending home of these workers when they exhibit symptoms typical of influenza becomes an important control measure. Illness in healthcare workers is likely to be under-reported because these people are highly motivated and typically do not want to burden co-workers, consider their roles important, and may not feel ?ill enough? to stay home. In addition, many paid healthcare workers may avoid using sick leave days that are combined with vacation time. Educational efforts explaining the seriousness of illness caused by influenza, its mode of transmission and nosocomial spread, as well as ways they can participate in control and prevention activities may be necessary to gain compliance with furlough of infected workers.


    Click to enlarge image


    Figure 2. Infection control measures for documented or suspected outbreaks. Adapted from www.cdc.govlncidod/hip/INFECT/flu_acute.htm


    Back to top

    Immunisation

    Annual influenza immunisation of high-risk persons and their contacts, including healthcare providers, is the primary means of preventing nosocomial influenza; however, vaccination has been substantially underused, particularly by healthcare workers.15
    Inactivated influenza vaccine in patients
    Increasing rates of pre-season immunisation among people in the community and healthcare providers would lessen the impact of influenza on the healthcare environment. Prevention of influenza has been hampered by inadequate vaccine use and by the fact that many high-risk patients do not develop a protective antibody response after immunisation. The Advisory Committee on Immunisation Practices (ACIP) of the Centers for Disease Control and Prevention (CDC) recommends annual immunisation vaccination of high-risk patients and their contacts.15 High-risk is currently defined as age 50 years or more, residence in a nursing home or other long-term care facility, chronic pulmonary or cardiovascular disorders, underlying illness that requires regular medical follow-up or hospitalisation, long-term aspirin therapy, and pregnancy entering its second or third trimester during an influenza season. Generally, in healthy young adults, influenza vaccine efficacy has ranged from 70?90% depending upon seasonal virus circulation.67 Among the US population over the age of 65, influenza vaccination rates have increased from 33% in 198968 to 63% in 1998.69 The greatest proportion of those vaccinated are non-Hispanic whites, and rates are lower among blacks and Hispanics.69 According to CDC, only 23% of adults younger than 50 years with high-risk conditions were vaccinated last year in the USA, far short of the 60% goal.15
    The effectiveness of vaccine in preventing influenza among the institutionalised elderly and those with chronic illness may be as low as 30?40%, but can be effective for reducing hospitalisations, complications, and deaths.70,71 One recent study suggested that an increased influenza vaccination rate among elderly patients in the community (to 65%) could have had an important effect (along with a significant increase in healthcare worker vaccination compliance rate) on risk for nosocomial influenza among patients in the local hospital.30
    Inactivated influenza vaccine in healthcare workers
    ACIP has also recommended annual influenza immunisation for healthcare workers since 1981 and has targeted providers in both hospitals and outpatient care settings as a high priority group since 1986.72 The goals of this strategy are to reduce the risk of patient influenza exposure and to ensure provision of essential community services, so that patient care is not disrupted. Vaccine acceptance by healthcare workers continues to be below recommended levels. One survey during the 1986?87 season at a large acute-care hospital found that among healthcare workers, immunisation rates were less than 10%.9 Also, 35% developed influenza-like illness during that influenza season, and 77% took care of patients while ill.9 Even in the setting of community-wide epidemics, several studies have reported that one-third or fewer hospital workers accepted influenza vaccine,73,74 and some studies have found lower rates of healthcare worker acceptance than among non-medical employees in such institutions.75 More recent surveys in the USA have found that approximately 30% of healthcare workers receive influenza immunisations.76
    Vaccine acceptance among healthcare workers
    Surveys of healthcare workers have identified various reasons for a lack of vaccine acceptance including inconvenience, necessity of obtaining written informed consent, concerns about vaccine effectiveness, concerns about vaccine side-effects including misconceptions that vaccine may cause influenza, and lack of understanding of the risks of acquiring infection and transmitting it to high-risk patients. One study found that the most frequently cited reasons for non-acceptance were fear of side-effects (35%), avoidance of medications (33%), reaction to vaccine in the past (24%), impression of low risk of acquiring influenza (18%), and dislike of shots (18%).75
    Specific occupational groups have different concerns, which remain to be fully elucidated. Better understanding of epidemiological risk factors and barriers to influenza immunisation should translate into better acceptance of vaccine. For example, a study at the University of Iowa74 found that physicians were significantly more likely to accept the vaccine than other occupational groups. Advancing age, prior absenteeism (as surrogate marker for underlying illness), higher socioeconomic status (salary level), and marriage were associated with increased vaccine uptake in various target groups.74 Another study found that the predictors of acceptance were prior receipt of influenza vaccine, age of at least 50 years, and knowledge that vaccine does not cause influenza.75
    Vaccine efforts need to be directed at all healthcare workers and other staff who deal directly with sick people. In this regard, the Iowa study concluded that professional support staff with patient contact, younger workers, those recently employed, and lower paid workers should be considered for targeted efforts at increasing immunisation.74 To maximise convenience and minimise disruption of usual clinical activities, the ACIP recommends using a mobile cart to deliver vaccine onsite to healthcare workers in the workplace, increased availability after hours, and follow-up immunisation programmes early in the course of recognised community outbreaks.15 At the University of Virginia, use of a chart, showing updated healthcare worker compliance rates with influenza vaccine, posted in frequented areas of the hospital, was partly responsible for increasing vaccination acceptance rates to nearly 70% (figure 1).30
    Vaccine efficacy
    Convincing evidence that influenza immunisation of healthcare workers would prevent nosocomial cross-infection would be a compelling argument in favour of healthcare worker immunisation. Two studies conducted in Scottish geriatric hospitals, one retrospective and the other prospective, provided evidence that healthcare worker immunisation was associated with reduced mortality during influenza season in the resident patients.8,77 However, it was unclear whether the reduction in mortality was due to prevention of influenza, since no data were presented regarding specific rates of influenza-related illness or pneumonia in patients or workdays lost in healthcare providers.8 Another study from the University of Virginia documented an association between a significant improvement in vaccine compliance among hospital employees and a significant decrease in nosocomial influenza among hospital patients over the same period.30
    Questions remain regarding the cost benefit of vaccinating healthy working adults because cost-benefit studies have not been consistent.78?81 A recent assessment of the societal economic impact of annual influenza vaccination of healthy working adults was conducted by Nichol.78 A mathematical model estimated that, vaccination prevented an average of 12 workday absenteeisms and three physician visits per 100 persons vaccinated per year. Vaccination also resulted in a net savings of almost $14 for each person vaccinated and was cost effective 95% of the time.78
    Few studies have focused specifically on healthcare workers. One double-blind, placebo-controlled trial found that vaccine efficacy based on serological evidence of infection was high in healthcare workers, mostly resident physicians and medical students.82 Compared with the cumulative influenza attack rate of 14% in the control group, protection against laboratory-documented influenza infection averaged approximately 90% for both influenza A and B virus infections. Reductions in days of reported febrile illness or days of absence were observed in vaccine recipients, but these were not statistically significant, perhaps due to relatively small samples sizes, intercurrent non-influenza illness, and a high degree of motivation to return to work. The finding that the mean number of reported febrile days exceeded the number of work absence days suggested that the affected healthcare providers reported for work during febrile illness82 (figure 3). Inactivated influenza vaccine is highly effective in preventing influenza infection in younger healthcare workers, reduces days of illness, and probably days of absenteeism.82


    Click to enlarge image


    Figure 3. Influenza vaccine in healthcare workers: cumulative outcomes.82


    Live-attenuated, intranasal influenza vaccine A
    live-attenuated, cold-adapted influenza vaccine (CAIV) administered intranasally is under active consideration by the United States Food and Drug Administration (USFDA) for use in adults younger than 65 years and children. One question regarding CAIV is whether it is as efficacious as the injected inactivated vaccine in adults. A 5-year controlled, blinded trial involving 5210 healthy adult, over 80% of whom were aged 18?65 years, compared the efficacy of inactivated vaccine to attenuated intranasal CAIV formulations containing two influenza A viruses. For H1N1 infection, the efficacy of intranasal vaccine was 85% in preventing culture-positive illness compared with 76% for injected vaccine. For H3N2 infection the efficacies were 58% and 76%, respectively. Inactivated vaccine significantly increased serum haemagglutination-inhibiting (HAI) antibody litres and appeared to be more protective against seroconversion and reported illness.83 However, this trial was limited by use of a ten-fold lower dose of CAIV virus than current formulations, administration by drops instead of spray (perhaps more effective), and, unlike the situation for injected vaccine components, by lack of frequent intranasal CAIV virus updates to assure a good antigenic match with the circulating strains of influenza. A recent study in experimental intranasally induced human influenza A and B found that the efficacy of a trivalent CAIV in preventing influenza following wild-type challenge was higher with CAIV-T (85%) than with trivalent inactivated vaccine (71%).84 A related question is whether vaccine recipients, symptomatic or not, are at risk of transmitting vaccine virus infection to hospitalised patients. In immunised adults the litres of recoverable virus in the nasal secretions have usually been much lower lhan Ihose needed lo induce infection in seronegalive children (103 50% tissue cullure infectious doses [TCID50]/mL) or adulls (105 TCID50/mL). Very rare instances of Iransmission of CAIV have been recognised from immunised children.
    Back to top
    Antiviral chemoprophylaxis

    Amantadine and rimantadine

    Two classes of antiviral agents, Ihe M2 ion-channel inhibitors (amanladine and rimanladine) and Ihe neuraminidase inhibitors (osellamivir and zanamivir) have proven efficacy in preventing influenza illness and infection. The firsl reported trial demonslraling the efficacy of antiviral chemoprophylaxis for hospital-acquired influenza was conducted in 197285 (figure 4). The trial was a non-blinded randomised comparison of oral amantadine with no treatment among patients admitted to either a general medicine or neurology ward during a recognised community epidemic of influenza A. Amantadine was given in a dose of 100 mg twice daily for the duration of hospitalisation. Four patients were admitted with active influenza. The efficacy of amantadine in preventing influenza illness was 100% compared with an attack rate of 11% in the group receiving no treatment. Almost all of the infections occurred in those with a low admission HI titre (less than or equal to 1/20) and only a few patients had received vaccine.85


    Click to enlarge image


    Figure 4. Prevention of nosocomial influenza with amantadine.85


    Emergence and transmission of drug-resistant influenza may occur rapidly with use of amantadine or rimantadine, specifically in institutions and households. Up to one-third of patients may shed resistant strains by 2?5 days after starting treatment for influenza56,86,87 Of note, among immunocompromised rimantadine recipients, 83% of those shedding virus for 3 days or more had resistant variants.56 Failures of drug prophylaxis and at least one case of fatal influenza due to probable nosocomial transmission of resistant virus have occurred in nursing home residents.88
    The optimal duration of antiviral chemoprophylaxis for a nosocomial outbreak is not certain. Randomised trials in nursing home populations, which employed short-term or long-term amantadine prophylaxis found that administration for 14 days, and 7 days beyond the last confirmed illness on the unit, appeared to be adequate.88 The duration of antiviral prophylaxis should be guided by ongoing surveillance on the specific ward or facility. Once drug use has been initiated, virologic testing of any new illnesses is helpful in guiding duration and to look for other nosocomial pathogens (eg, respiratory syncitial virus [RSV]). Persistent isolations of virus with new cases suggest circulation of drug-resistant variants or new introductions. For example, in nursing homes, outbreaks of parainfluenza virus, RSV, and influenza B virus have been documented during periods of community influenza A activity or there have been out-of-season outbreaks not temporally associated with community activity.89
    Amantadine and rimantadine can cause central nervous system (CNS) and gastrointestinal side-effects. The incidence of CNS side-effects seems to be greater for those taking amantadine. One nursing home study evaluated CNS toxicity among patients receiving amantadine or rimantadine for influenza A prophylaxis during the 1997?1998 influenza season.90 Amantadine was associated with a significantly higher rate of CNS toxicity than rimantadine (18?6% vs 1?9%, p<0?01) and more patients discontinued amantadine (17?3% vs1?9%, p<0?001). Confusion was the predominant symptom associated with CNS toxicity and after multivariate analysis, male sex, reduced creatinine clearance, and amantadine use were all significant risk factors for CNS toxicity.90
    Back to top

    Oseltamivir and zanamivir

    The neuraminidase inhibitors (zanamivir and oseltamivir) have demonstrated safety and efficacy in prevention of both influenza A and B in community and family-based studies.45,91 Oseltamivir may be given orally; however, because of poor bioavailablity, zanamivir must be given by inhalation. One controlled trial of inhaled zanamivir for prevention of influenza in families with an index case of illness demonstrated that zanamivir provided protection against influenza A and B with a 79% reduction in the proportion of families with at least one affected contact.45 Also, importantly, no zanamivir-resistant variants were recovered among index cases who received zanamivir. The protective efficacy of long-term use of oseltamivir during an influenza season was 74% for influenza-like illness and 87% for culture-proven disease.91 When used for post-exposure prophylaxis in families, oseltamivir (75 mg once daily for 7 days) was 90% effective.92 Clinical isolates with reduced susceptibility to zanamivir and oseltamivir after treatment have been reported uncommonly;Roche Laboratories 2000 70 /id but data are limited regarding possible transmission of these strains and their ultimate clinical impact.
    The use of zanamivir and oseltamivir during acute-care outbreaks of influenza has not been reported, but evidence exists regarding effective influenza prevention in other institutional settings. One randomised, unblinded study compared zanamivir chemoprophylaxis with rimantadine during an influenza A nursing home outbreak and to no prophylaxis during a subsequent influenza B outbreak.93 During the prophylaxis period for the influenza A outbreak, only one laboratory-confirmed case of influenza A occurred in a volunteer taking rimantadine and during the prophylaxis period for the influenza B outbreak, only one laboratory- confirmed case of influenza B occurred in a volunteer not receiving prophylaxis.93 Even though this study was small with a low overall influenza attack rate, it demonstrated that zanamivir was comparable to the standard of care at the time. A follow-up, blinded trial study compared 2-week prophylaxis with oral rimantadine (100 mg daily) with inhaled zanamivir (10 mg daily in nursing homes with documented influenza A outbreaks) and found that zanamivir was associated with 61% additional protective efficacy compared with rimantadine, in part because of failures of rimantadine prophylaxis due to drug-resistant strains.48 In a highly immunised nursing home population, seasonal prophylaxis with oral oseltamivir 75 mg once daily was well-tolerated and 92% preventive against influenza compared with placebo.94 These newer antiviral agents are highly effective for prevention but more expensive than M2 inhibitors, and other than documented influenza B outbreaks, their role in prevention of nosocomial outbreaks in acute-care hospitals remains to be determined. Given the significant healthcare effects associated with nosocomial influenza outbreaks, CDC recommends that institutions consider prophylaxis for exposed patients, unvaccinated personnel, and those vaccinated less than 2 weeks before exposure. When choosing which antiviral to use for prophylaxis, side-effects, cost, and influenza strain and susceptibility pattern should be considered.
    Back to top
    Search strategy and selection criteria

    Medline searches of English language publications from 1966?2001 on the major topic headings: influenza, cross infection, nosocomial influenza, morbidity, mortality, death, patient outcome, influenza vaccine, infection control, cost, cost benefit analysis, prevention, and treatment (and combinations thereof) as well as personal Reference Manager files and bibliography lists of textbook chapters were used as the database for this review. References were selected for use if they addressed the impact, transmission, prevention, or control of influenza in the acute-care hospital, or the efficacy and recommendations for influenza vaccine and chemoprophylaxis.




    Back to top

  • #2
    Figure 1


    Figure 1. Poster displayed at University of Virginia showing proportion of UVA healthcare workers receiving influenza vaccine.

    Comment


    • #3
      Table 1


      Table 1. Description of reported nosocomial outbreaks in the acute-care setting

      ICU= intensive care unit; WBC=white blood cells; HCW=healthcare workers

      <!--start ce:table-footnote=-->
      *<!--start ce:note-para=-->Description of nosocomial influenza infections during a 1 month period of increased influenza activity in the community

      Comment


      • #4
        Figure 2


        Figure 2. Infection control measures for documented or suspected outbreaks. Adapted from www.cdc.govlncidod/hip/INFECT/flu_acute.htm

        Comment


        • #5
          Figure 3


          Figure 3. Influenza vaccine in healthcare workers: cumulative outcomes.<!--start ce:cross-ref=--><!--start ce:sup=-->82<!--end ce:sup--><!--end ce:cross-ref-->

          Comment


          • #6
            Figure 4


            Figure 4. Prevention of nosocomial influenza with amantadine.<!--start ce:cross-ref=--><!--start ce:sup=-->85<!--end ce:sup--><!--end ce:cross-ref-->

            Comment

            Working...
            X