Antivirals for Pandemic Influenza: Guidance on Developing a Distribution and Dispensing Program (2008)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

2 Antiviral Effectiveness, Safety, and Supply T wo neuraminidase inhibitors, oseltamivir and zanamivir, are the antivirals currently being stockpiled to respond to an influenza pandemic. Oseltamivir, sold under the name Tamiflu, is manu- factured by Roche and Gilead and is available in capsule form and as a suspension for pediatric use. Zanamivir, sold as Relenza, is manufactured by GlaxoSmithKline and is delivered as a powder via inhaler. The two drugs are effective when used for post-exposure and extended prophy- laxis for seasonal influenza, and when used for treatment to shorten the duration of both symptoms and viral shedding (European Medicines Agency, 2005). In this chapter, the committee examines current information about the effectiveness and safety of neuraminidase inhibitors, and implications of supply issues, and emergent effectiveness and safety data for planning and implementation of an antivirals program. Supply issues include the source and the adequacy of existing drug stockpiles to meet stated plan- ning purposes (treatment versus treatment and prophylaxis); the cost of purchasing and storing the drugs; and the drugs’ shelf-life. Many of the relevant stakeholders (Association of State and Territorial Health Officials [ASTHO], state public health agencies, large health care delivery systems) consider these issues crucial to examine in the planning process and necessary to address long before an influenza pandemic begins (ASTHO, 2006; Skivington and Koscove, 2008). 19 

20 ANTIVIRALS FOR PANDEMIC INFLUENZA EFFECTIVENESS AND SAFETY The currently approved treatment for seasonal influenza with neur- aminidase inhibitors requires a 5-day regimen (one 75 mg capsule two times per day of oseltamivir, or two inhalations two times per day of zanamivir), while prophylaxis for exposure to seasonal influenza virus requires a 10-day regimen (one capsule per day or two inhalations per day). Randomized trials have demonstrated efficacy when neuramini- dase inhibitors, in particular oseltamivir, were used for prophylaxis after exposure in both family and institutional settings and when used for 6 weeks as seasonal prophylaxis in nursing homes (Moscona, 2005; Hayden and Pavia, 2006). Data from a study of hospitalized adults demonstrated a strong association between oseltamivir use and decreased mortality (adjusted odds ratio 0.21) (McGeer et al., 2007). The two major gaps in the research on neuraminidase inhibitor effectiveness include lack of high- quality data in high-risk persons and lack of adequate data for H5N1. Although neuraminidase inhibitors have not been studied or approved for prolonged use beyond 6 weeks, protecting vulnerable health care and other workers during a pandemic may require prophylaxis over several weeks or longer. Pandemic duration likely will depend on the success of non-pharmaceutical interventions in “flattening” (CDC, 2007c) the pan- demic curve. This refers to decreasing and extending the plotted course of a pandemic, which in the absence of effective interventions the first wave of a pandemic is expected to have an earlier, more severe peak in the num- ber of cases per week. Modeling and data from the 1918 pandemic suggest that effective non-pharmaceutical interventions would delay and decrease the height of the peak, but result in a longer duration of the pandemic wave (IOM, 2006). However, models have limitations, and the vastly dif- ferent socio-cultural setting of historic data must be acknowledged. Neuraminidase inhibitors are effective in treating influenza and are currently used to prevent some cases during seasonal influenza outbreaks. Randomized trials, conducted primarily among relatively healthy chil- dren and adults, have demonstrated decreased duration of symptoms, disability, antibiotic use, and decreased time to return to normal function after treatment (Moscona, 2005). However, treatment is most effective when given as early as possible after symptoms develop, and its effective- ness diminishes markedly after 48 hours. Data are limited on the efficacy in high-risk populations. The available data suggest decreased hospital- ization and mortality rates with treatment; however these are derived   Oseltamivir adult dose: 75 mg bid for 5 days for treatment, 75 mg qd for 10 days for prophylaxis (Roche Pharmaceuticals, 2008); oseltamivir pediatric dose (ages 1–12 years): ≤33 lbs, 30 mg/bid/5 days; >33 lbs to 51 lbs, 45 mg/bid/5 days; >51 lbs to 88 lbs, 60 mg/bid/5 days; >88 lbs, 75 mg/bid/5 days (add: same dose qd/10 days for prophylaxis). 

ANTIVIRAL EFFECTIVENESS, SAFETY, AND SUPPLY 21 from pooled analysis and observational studies. One study (McGeer et al., 2007) suggests a survival benefit may be seen among hospitalized patients even if treatment is started later than 48 hours from symptom onset. In past pandemics and in seasonal influenza outbreaks, morbidity and mortality have been high among infants and pregnant women. There currently are no pharmacokinetic, efficacy, or safety data for oseltamivir in infants under 1 year of age or in pregnant women (CDC, 2007b). Because very little zanamivir is absorbed systemically, it has been considered an alternative for pregnant women (DHHS, 2005), but this is not based on adequate data. If response to a pandemic necessitates use of antivirals in populations for which the drugs have not been indicated, careful follow- up or monitoring of adverse events would be needed (see Chapter 4 for additional discussion). Use of antivirals in the context of a pandemic will test what is known about both their effectiveness and safety. It is not known if the effective- ness of neuraminidase inhibitors against the pandemic strain will be simi- lar to their effectiveness when used during seasonal influenza outbreaks. Wide-scale use, as well as overuse or misuse, could lead to both increased resistance and the emergence of newly identified, or at least greater num- bers of known, adverse events. It also is possible that the currently recom- mended dosage and duration of treatment will need to be reconsidered during a pandemic. In fact, there is some evidence from studies in animals that high levels of viral replication may require higher doses and/or lon- ger duration of treatment for effectiveness (WHO, 2008). Safety Neuraminidase inhibitors are generally well tolerated and have a favorable risk–benefit profile. However, serious adverse events reported in postmarketing use in individuals who took oseltamivir include skin reactions, neuropsychiatric events, and pediatric deaths (the latter two largely reported from use in Japan) (FDA, 2005; Roche Pharmaceuticals, 2008). Gastrointestinal disturbances are reported in about 10 percent of users. Adverse events experienced by people who took zanamivir include headaches, gastrointestinal problems, respiratory problems, dizziness, and musculoskeletal symptoms (GlaxoSmithKline, 2007). Postmarketing adverse events also include allergic, cardiac, neurologic, and skin reac- tions, as well as bronchospasm. Oseltamivir and zanamivir have been on the market for more than 8 years and more is known about their safety profiles now than at the time of approval. However, wide-scale and prolonged use of these drugs in populations who were not well studied may lead to the recognition of rare but serious adverse events. Moreover, with the widespread use 

22 ANTIVIRALS FOR PANDEMIC INFLUENZA anticipated during a pandemic, reports of possible adverse events will be much more likely and will need to be rapidly evaluated. A telling example is found in reports of neuropsychiatric adverse events from Japan where oseltamivir is used much more widely than in the United States in treat- ing seasonal influenza (FDA, 2007b). Despite two careful reviews by Food and Drug Administration (FDA) staff and FDA advisory committees, it has not been possible to date to determine if the neuropsychiatric events are causally associated with oseltamivir. Appropriate systems will be needed to gather, analyze, interpret, and act on the drug safety information collected (see discussion in Chapter 4). Given the challenging circumstances of a public health emergency, it is important to plan in advance for assessment of and communication about adverse events and for possible changes in antiviral dispensing strategies to keep the public’s confidence in the public health effort (see discussion in Chapter 3). Resistance The effectiveness of adamantanes, which have been marketed for three decades, has become so limited by the spread of resistant viruses that federal and international public health authorities advise against their use in treating seasonal influenza (CDC, 2006a). Adamantane-resistant virus does not appear to have compromised infectiousness or transmissi- bility. The neuraminidase inhibitors are considerably newer drugs. Before 2008, rates of resistance to neuraminidase inhibitors observed in clinical trials and population surveillance were low. In clinical trials, the rate of emergence of resistance to neuraminidase inhibitors was about 1 percent among adults and approximately 5 percent among children (European Medicines Agency, 2005; Aoki et al., 2007). Among samples submitted to the Neuraminidase Inhibitor Susceptibility Surveillance Network in 2003–2004, fewer than 1 percent were resistant. Animal studies of resistant virus suggested that many, but not all, strains had compromised transmis- sibility and growth characteristics (or “fitness”). These clinical and labora- tory data led to assumptions that resistance to neuraminidase inhibitors would emerge and spread slowly. In the 2006–2007 influenza season, 0.5 percent of viral strains isolated (776 of 2121 H1N1) were oseltamivir resistant (Klimov, 2008). However, during the 2007–2008 influenza season, there has been some evidence of an increase in the incidence of oseltamivir-resistant strains of influenza A(H1N1) detected in Europe, Australia, Hong Kong, and North America (EMEA, 2008; Reuters, 2008). For example, 14 percent of 437 H1N1 isolates in Europe were oseltamivir resistant (Lackenby et al., 2008). This suggests that transmissible and virulent neuraminidase inhibitor–resistant viruses 

ANTIVIRAL EFFECTIVENESS, SAFETY, AND SUPPLY 23 may spread more readily than previously thought. It is important to note, however, that in Europe use of oseltamivir for treatment of seasonal influenza is limited, thus the emergence of resistance may not be related to use. Japan, with its higher rates of seasonal oseltamivir use has not yet reported an increase in resistant strains. The implications of increasing antiviral resistance for pandemic planning are unclear, but troubling. The factors influencing the rate of antiviral resistance are complex. Based on what is known about the emergence of antiviral resistance in HIV, potential factors driving the emergence of neuraminidase inhibitor– resistant influenza include the structure of the circulating neuraminidase, the relationship of the susceptibility of the pandemic strain and the drug levels among the patients, and the degree of inappropriate use or partial prophylaxis in the presence of circulating resistant virus. Perhaps the most important drivers will be the number of courses of drug that are used and adherence to the drug regimen. Widespread use for post-exposure and seasonal prophylaxis as well as treatment may create selective pressure for the emergence of resistance (Lipsitch et al., 2007). However, the rate of emergence and spread of resistant viruses must be measured against the time it takes to produce an effective vaccine. If neuraminidase inhibitors remain largely effective until a well-matched vaccine can be deployed, there will be a large net benefit despite the emergence of resistance. Con- versely, delays in vaccine availability could coincide with rapid develop- ment of neuraminidase inhibitor resistance, presenting a grave challenge to effective pandemic response. Mathematical modeling of resistance in the context of treatment and prophylaxis provides some insight, albeit with a high degree of uncer- tainty. Lipsitch and colleagues (2007) modeled the predicted impact of four strategies for antiviral use on the number of cases and the emergence of resistance: no antivirals; antivirals for treatment only; antivirals for household prophylaxis without treatment; and antivirals for treatment and household prophylaxis. They predicted that use of antivirals exclu- sively for treatment led to the least emergence of resistance. Exclusive use for household and seasonal prophylaxis eventually led to significant emergence of resistant virus in the model, but only after some lag time. However, Lipsitch and colleagues predicted that combined use for both treatment and prophylaxis led to the most widespread and rapid emer- gence of resistant virus. SUPPLY The committee believes that after resistance-modified effectiveness, the second core issue in identifying the most appropriate strategies for dispensing the medication is the quantity of available antiviral drugs. 

24 ANTIVIRALS FOR PANDEMIC INFLUENZA Shelf-life, manufacturing capacity, cost, and storage requirements are some of the important variables linked with quantity of antiviral supply. Existing Stockpiles The Division of Strategic National Stockpile (SNS) at the Centers for Disease Control and Prevention (CDC) holds a large number of courses of neuraminidase inhibitors, being purchased in a ratio of approximately 80–85 percent oseltamivir to 15–20 percent percent zanamivir. (The SNS also includes a small number of courses of rimantadine.) The total goal for the SNS is 81 million courses of antivirals, with 50 million being stockpiled by the federal government and 31 million courses of neuraminidase inhibitors intended for procurement by states through the federally subsidized purchasing program. Of the 81 million, 6 million are intended for containment of the pandemic, and 75 million (in federal SNS and state stockpiles combined) are to be used to treat 25 percent of the population (see Table 2-1). As of March 7, 2008, the federal stockpile was nearly complete (goal of 50 million), and state stockpiles were two- thirds complete. (For more information on state plans, see Appendix B.) Federal government documents are not yet clear on whether the goal of antiviral use will be treatment or a combination of treatment and pro- phylaxis. The Homeland Security Council National Strategy for Pandemic Influenza states that “current plans propose using antiviral medication stockpiles only for treatment once a pandemic is underway. Prophylac- tic use of antiviral medications will be reserved for initial containment efforts and other highly select circumstances” (Homeland Security Coun- TABLE 2-1  Strategic National Stockpile Antiviral Goal and Status Status Status March Goal July 2007 2008 Federal Strategic 50 million courses (6 million 36 million 49.9 National Stockpile of these are intended for million domestic containment) Federally subsidized, 31 million courses 12 million 21.7 state-purchased state million stockpiles Total 81 million courses 48 million 71.6 million SOURCES: DHHS, 2007a,c; A. Patel, personal communication, March 10, 2008. 

ANTIVIRAL EFFECTIVENESS, SAFETY, AND SUPPLY 25 cil, 2006:106). The Department of Health and Human Services (DHHS) Pandemic Influenza Plan (DHHS, 2005) provides “recommendations . . . on the distribution and use of antiviral drugs for treatment and prophylaxis throughout the pandemic phases” (DHHS, 2005:11; see Box 2-1 for a list of the groups identified to receive prophylaxis in the 2005 plan). The DHHS draft proposed guidance on antiviral use similarly describes use of antivirals for treatment and for a potential range of prophylaxis activities (DHHS, 2007a). DHHS references to the stockpile, however, describe it as “antiviral treatment courses for 25 percent of the U.S. population or 81 million treatment courses” (Vanderwagen, 2007), and the DHHS Secretary’s Pandemic Planning Update IV (DHHS, 2007c) refers to the stockpile as “81 million treatment courses,” which “include 6 million treatment courses set aside for the early stages of an emerging pandemic.” Clearly, the stockpile of 75 or 81 million courses is not suf- BOX 2-1 Priority Groups to Receive Antiviral Treatment and Prophylaxis Identified in the 2005 DHHS Pandemic Influenza Plan (based on the recommendation of the National Vaccine Advisory Committee) • Patients admitted to hospital • Health care workers (HCWs) with direct patient contact and emergency medical services (EMS) providers • Highest-risk outpatients—immunocompromised persons and pregnant women • Pandemic health responders (public health, vaccinators, vaccine and antiviral manufacturers), public safety (police, fire, corrections), and government deci- sion makers • Increased-risk outpatients—young children 12–23 months old, persons ≥65 years old, and persons with underlying medical conditions • Outbreak response in nursing homes and other residential settings • HCWs in emergency departments, intensive care units, dialysis centers, and EMS providers • Pandemic societal responders (e.g., critical infrastructure groups as defined in the vaccine priorities) and HCWs without direct patient contact • Other outpatients • Highest-risk outpatients • Other HCWs with direct patient contact SOURCE: DHHS, 2005:D-10. 

26 ANTIVIRALS FOR PANDEMIC INFLUENZA ficient to both treat 25 percent of the population and provide any level of prophylaxis (at the low end of the range, post-exposure prophylaxis for certain health care and emergency workers; at the high end household post-exposure prophylaxis as described in the Community Mitigation Strategy) (CDC, 2007a). Providing any level of prophylaxis with existing drug supplies would require limiting the proportion to be used for treat- ment. Alternately, if the intention is to treat 25 percent of the population and provide some level of prophylaxis, a stockpile sufficient to meet those goals will be needed. Recommendation 2-1: The committee recommends that the fed- eral government clarify the national goals for antiviral use in an influenza pandemic. If these goals include treatment of all antici- pated cases and a level of prophylaxis, fiscal appropriations will be needed to expand the national stockpile to meet these goals.  Most states have purchased some or all of their portion of the feder- ally subsidized stockpile (enough to treat 25 percent of a state’s popula- tion) (ASTHO, 2007; A. Patel, personal communication, March 10, 2008). State rationale for antiviral policy differences aside, the variations in availability of antivirals for treatment across the country may result in some geographic inequities, with populations residing in one state, but not another, having potential access to antiviral treatment. Further, sea- sonal shifts in population density may need to be considered. The pat- tern of disease spread across the United States may add another layer of complexity. Ensuring������������������������������������������������� equitab����������������������������������������� ������������������������������������������������ le access ������������������������������� to household post-exposure pro- phylaxis (if this were a national goal for antiviral use) ������������������� within states ����� will depend on addressing��������������������������������������������������� access barriers that may be faced by cultural and linguistic minorities and other vulnerable populations. Adequacy of Supply In thinking about the available public stockpile (at federal, state, and to a lesser extent, local levels) and the important issues listed above, the committee used three simple potential scenarios to think about antiviral dispensing. The scenarios assume different sizes of antiviral stockpile each of which is adequate for a certain level of dispensing and so, for some, but not all, dispensing goals. Scenarios A and C suggest two extremes of possible antiviral use scenarios, assuming some level of antiviral stockpil-   Recommendations are numbered by the chapter where they occur and their order in the chapter. There are no recommendations in Chapter 1, so Recommendation 2-1 is the first recommendation in the report. 

ANTIVIRAL EFFECTIVENESS, SAFETY, AND SUPPLY 27 ing (i.e., not zero). Scenario B describes a level of stockpiling somewhere between A and C. However, the committee recognizes that decisions about the amount to stockpile and the goals of antiviral use are complex and informed by many other considerations, both related and unrelated to pandemic influenza planning. Antiviral distribution and dispensing activities would vary with sup- ply available (or conversely, supply would depend on the antiviral use goals selected), community circumstances, and emerging information about the severity and the epidemiology of the disease (such as what age group is most affected, who is at increased risk of death, who is transmit- ting the virus, etc.). With regard to severity, DHHS plans provide a 1–5 rating of severity, with <0.1 percent case fatality rate as Category 1 and ≥2.0 percent as Category 5 (CDC, 2007a). Scenario A In this scenario, the antiviral stockpile is 81 million courses, an amount that would treat a considerable proportion of cases, or provide a considerable level of prophylaxis (depending on specific objectives), or satisfy both goals only in a limited way. Use of antivirals for prophylaxis only has been considered at least in modeling of resistance and effective- ness (e.g., McCaw and McVernon, 2006; Lipsitch et al., 2007), but short of such use to contain the pandemic in the early days of a U.S. outbreak, limiting dispensing to prophylaxis may imply denying patients the only potentially effective treatment. Based on some of the estimates provided in the November 6, 2007, DHHS draft proposed guidance, the groups requiring prophylaxis for occupational exposure would require most of the stockpile for several courses of outbreak (or seasonal) prophylaxis, leaving little or nothing for treatment. Reserving the antivirals for treatment alone is not an option as it would leave a vast proportion of health care and emergency services personnel with certain exposure unprotected and vulnerable to becom- ing incapacitated by illness. Even if the intent was to provide only post- exposure prophylaxis to 10.7 million health care workers and emergency services personnel (using figures from the DHHS draft proposed guid- ance, which estimates that one-third of health care workers would not have direct contact with infected individuals), 2 or 3 courses for each of those workers would amount to a total of 21.4 or 32.1 million courses, which would considerably limit the supply available to treat cases.   As noted earlier, this is the current goal for the Strategic National Stockpile and sub- sidized state stockpiles, and this total includes 6 million courses intended to be used for containment. 

28 ANTIVIRALS FOR PANDEMIC INFLUENZA The best that could be achieved with this level of antivirals would be to develop dispensing strategies to treat only some cases and provide prophylaxis to a narrowly defined subset of those with occupational expo- sure. Once the epidemiology and virology of the disease and its agent became known, dispensing strategies could target the proportion of cases more likely to require hospitalization and more likely to die, in order to allow for the use of a portion of the stockpile for post-exposure prophy- laxis of some health care workers and emergency personnel (see Chapter 4 for additional discussion). Accurate diagnosis of pandemic influenza would be crucial to ensure that antivirals are used only for confirmed cases; this may mean that remote diagnosis may be less desirable, but that could lead to later diagnosis, perhaps past the 48-hour window when antiviral efficacy may be highest. Due to the limited antiviral supply, no household post-exposure prophylaxis would be provided in this scenario, so dispensing sites and related activities for exposed individuals without occupational risk would not be needed. The risk of resistance may be smaller than in a scenario with broader prophylaxis, and resistance would spread more slowly because some of the drug courses would be used for treatment, which is less likely than prophylaxis to lead to resistance (Lipsitch et al., 2007). A pandemic of any level of severity greater than seasonal influenza would cause some level of social and economic disruption due to higher than usual hospitalization and death rates, but a severe pandemic in this scenario of a modest antiviral stockpile would involve making the most difficult decisions, and thus, present the greatest need for a pre-developed, widely understood ethical framework. In the event that emerging infor- mation about the pandemic strain results in changing the dosage or dura- tion of antiviral use, this would place additional demands on the limited supply of antivirals. Scenario B Somewhere between scenarios A and C, the committee envisioned an antiviral stockpile sufficient to provide treatment to most or all cases and prophylaxis to defined groups with occupational risk. Well more than twice the existing goal of 81 million would be needed to treat 25 percent of the population and provide outbreak and post-exposure prophylaxis to broadly defined groups with occupational exposure. The draft pro- posed DHHS guidance on antivirals identifies about 8.7 of the 13 million   This is assuming that 25 percent is sufficient, but for reasons described elsewhere in this report, such as change in dosage or higher attack rate, a greater number of courses may be needed. 

ANTIVIRAL EFFECTIVENESS, SAFETY, AND SUPPLY 29 TABLE 2-2  Prioritized Strategies for Antiviral Drug Use from November 2007 DHHS Draft Proposed Guidance Estimated number of Population to receive prophylaxis antiviral courses needed } Initial pandemic outbreaks overseas and in the United States 6 million Exposed travelers entering the United States early in a pandemic Persons with pandemic influenza illness (outbreak 79 million and post-exposure) Health care and emergency services workers 103 million Outbreak control in closed settings (e.g., nursing 5 million homes) Immunocompromised and not candidates for 2 million vaccine Unique and specialized infrastructure workers 2 million Household contacts of cases 88 million The summary of the proposed guidance, dated November 20, 2008, revises the preliminary position on household prophylaxis: “No national recommendation is made at this time for PEP [post-exposure prophylaxis] of household contacts of an influenza case or for workers in sectors other than healthcare and emergency services.” Total estimated number of courses for treatment 285 million and prophylaxis Total excluding household post-exposure prophylaxis 197 million SOURCE: DHHS, 2007a,b. health care workers and the 2 million emergency services personnel in the United States, as needing outbreak (or seasonal) prophylaxis (DHHS, 2007a). A population of 10.7 million requiring 8 courses of antivirals (DHHS’s estimate for 12 weeks of protection) would total 85.6 million courses. The draft proposed guidance also estimates needing approxi- mately 4 courses of post-exposure prophylaxis for each of the remaining 4.3 million health care workers not identified as needing outbreak pro- phylaxis, for a total of 17.2 million courses. Combined, these preliminary estimates for courses needed to provide prophylaxis to health care and emergency services personnel total nearly 103 million (see Table 2-2 for additional populations and estimates provided in the summary draft 

30 ANTIVIRALS FOR PANDEMIC INFLUENZA proposed guidance documents). The capacity to manufacture approxi- mately 80 million courses of oseltamivir (plus some additional amount of zanamivir) per year may make possible an increase in the stockpile in the intermediate term (1–2 years) to provide the level of treatment and prophylaxis described above. Scenario C In this scenario, a large amount of courses is available for a wide range of uses. One figure to illustrate this is 285 million courses (DHHS, 2007a), based on the estimates provided in the DHHS draft proposed guidance. An even higher total would ensure one course of treatment is available for all people in the United States at the time of the pandemic, a combina- tion of post-exposure or seasonal prophylaxis for all health care workers and emergency services personnel, and prophylaxis for a large number of exposed household contacts (again using DHHS 2007 estimates). In this scenario, sufficient antivirals are available for treatment and for expanded prophylaxis (both post-exposure and outbreak or seasonal, the latter requiring multiple courses of prophylaxis to protect from ongo- ing occupational exposure) such as that described in the DHHS draft proposed guidance. If planning for the 285 or 197 million antiviral courses (totals with and without post-exposure prophylaxis for household con- tacts, who would require a large portion of an antiviral supply) estimated in the draft proposed guidance, this scenario would require approxi- mately 2.5 to 3.5 times the SNS goal of 81 million courses. The committee is not aware of plans to bring the total of federally stockpiled antivirals to that level, and as noted above, there are other considerations that inform stockpiling decisions. Unlike a shortage scenario A, greater availability of antivirals could mean lessened concern about securing antiviral stockpiles and perhaps decreased potential for fraudulent attempts to obtain and sell antivirals. Scenario C would require the most diverse array of dispensing sites given the breadth of the prophylaxis being offered. The potential for more rapid development resistance may be greatest in this scenario, and the conse- quences would be most concerning in a severe pandemic. A rapid increase in resistance, facilitated by extensive prophylaxis, could undermine one of the main objectives in using antivirals: to “buy” time for the development of a well-matched vaccine. Wide-scale use in a less severe pandemic could offer early opportunities for safety signals to emerge and to be captured, assuming that health care providers are more likely to be able to report adverse events. 

ANTIVIRAL EFFECTIVENESS, SAFETY, AND SUPPLY 31 Additional Observations About Antiviral Supply As noted on previous pages, pandemic influenza planners hope that antivirals will be an effective tool in reducing death and hospitalizations until a vaccine becomes available. Decision makers have been and will need to continue to consider a wide range of factors in determining what level of antiviral stockpiling is desirable, based on the goals of an anti- viral dispensing program and the size of the populations to be targeted. Production capacity is one of those factors, and for oseltamivir specifi- cally, it has increased to approximately 80 million courses per year (U.S. production alone), thanks in part to the manufacturer’s ability to synthe- size shikimic acid (Roche Pharmaceuticals, 2006). It is important to note an additional factor in planning for antiviral supply: the two antivirals thought to have the greatest potential for use in a pandemic remain on patent until 2016 so no generic versions are available. The goal of stockpiling enough to treat 25 percent of the population may have been established by assuming that not all who develop influ- enza (with a 30 percent attack rate) will present for treatment within 48 hour after symptom onset, and also perhaps that not all cases will require some kind of medical care. Even so, there is reason to believe that addi- tional demands will be placed on the antiviral supply. The 2005 DHHS Pandemic Influenza Plan includes patients hospitalized with pandemic influenza on the list of priority groups to receive antivirals, and acknowl- edges that hospitalized patients would include individuals presenting for care late after falling ill. Further, if the attack rate is higher than the 30 percent planning assumption, additional drugs may be needed for treat- ment. Also, the absence of a reliable, rapid, point-of-care diagnostic test on the one hand, and likelihood of some level of remote diagnosis (telephone or web-based) on the other hand, would lead to treating patients who do not have pandemic influenza. With a larger quantity of available medication, a broader range of groups could be offered prophylaxis, and there would be less need to prioritize among them. However, greater financial and other resources also would be needed to purchase, distribute, and dispense the antivirals. Also, there would be opportunity costs of allocating resources to an inter- vention that is potentially ineffective or only partially effective against a pandemic viral strain, as well as the possibility of rapid development of resistance with widespread use. There may be other types of interven- tions that could be considered, for example, public health agencies and hospitals may weigh investing in antivirals against purchasing additional personal protective equipment or ventilators. Further, antivirals represent one intervention for only one kind of threat to the public’s health in a context of chronic underfunding for most public health agencies and pro- 

32 ANTIVIRALS FOR PANDEMIC INFLUENZA grams and inadequate federal budget allocations to address other critical areas of disease prevention (IOM, 2003; APHA News, 2008). Shelf-Life The expiration dating of oseltamivir capsules was recently extended from 5 to 7 years, on FDA approval of Roche’s supplemental new drug appli- cation in December 2007 (Duffy, 2007). Although FDA’s approval applies to oseltamivir capsules manufactured after 30 days from the approval, this extension has also been applied to state oseltamivir procurements. As an example, drugs purchased by a state in 2006 with an expiration dating of 5 years would not expire in 2011, but be considered viable until 2013. The 7-year expiration dating of oseltamivir in government stockpiles currently does not apply to current commercial product, that is, sold via commercial wholesalers or dispensed by pharmacists to individuals purchasing on their health care provider’s prescription. Aside from the change in expiration dating of oseltamivir, it is impor- tant to note that federal antiviral stockpiles are included in an FDA- overseen Shelf-Life Extension Program (SLEP) (FDA, 2007a) that is “ori- ented towards the testing of ‘military significant’ products, those that are either military-unique, possessing no commercial (non–Department of Defense) market, or drugs the Federal Government procures in such large quantities, for pre-positioned stocks, that vendors are unwilling to accept them for credit upon expiration” (DoD, 2007:2). Although states are provided a 25 percent subsidy to purchase antivirals as part of the SNS program, stockpiles built by states with partial federal funding are not included in the federal Shelf-Life Extension Program (ASTHO, 2008). The National Strategy for Pandemic Influenza: Implementation Plan (released May 2006) tasked DHHS and the Departments of Veterans Affairs and Defense with exploring “the possibility of broadening SLEP [the Shelf-Life Extension Program] to include equivalently maintained State stockpiles” and with providing an answer “within 6 months” (HSC, 2006). The committee has learned that at the time of this writing, develop- ment of a SLEP for non-federal stockpiles continues to be under discussion in DHHS (D. Wawrose, personal communication, February 20, 2008). The committee found a modest amount of information about SLEP and nothing about the feasibility, cost, and other barriers of extending the program to properly maintained non-federal stockpiles, including state and perhaps even some private-sector stockpiles. Thirty-one million of the 81 million course goal for the SNS would be held in state stockpiles. This raises the possibility that in the absence of a pandemic, large amounts of the drugs will expire and need to be discarded, an outcome that could be avoided or delayed if this considerable proportion of the nation’s govern- 

ANTIVIRAL EFFECTIVENESS, SAFETY, AND SUPPLY 33 ment stockpile were included in SLEP. The cost of antivirals, especially given the potential of expiration before a pandemic occurs, also is an issue for states and private-sector entities seeking to build their own stockpiles. Regimens of antivirals cost approximately $70–100 for a 10-capsule treat- ment course when purchased commercially (compared to approximately $20 per course for government purchases for the SNS) (A. Patel, personal communication, February 28, 2008). Considering the three scenarios described above also requires noting that the shelf-life of the antivirals in the SNS held by the federal gov- ernment is unknown—information is not being disclosed about when antivirals currently stockpiled are expected to expire (according to the non-disclosure agreements signed by states, information about the SNS has the status of Sensitive But Unclassified, and as such, is not subject to the Freedom of Information Act). Furthermore, stocks of antivirals in the SNS or in state stockpiles may not be rotated to allow use of a portion of the drugs during seasonal influenza and their replacement with fresh antivirals—in fact, rotating is not permitted. Rotating stocks of antivirals may pose logistic challenges, and the amount of drug used in the United States for treatment during seasonal influenza outbreaks may be too small to enable complete rotating of stocks. Despite these issues, the fact that rotating stocks is not allowed precludes the possibility of considering alternatives that could have economic and preparedness benefits. The committee believes that use of the SNS may be constrained in ways that are counterproductive to effective preparedness and, ultimately, to an effective response. The shelf-life of antivirals for pandemic influenza is an important factor and an economic barrier because a short shelf-life would require discarding previously purchased drugs and buying new ones. There is a need to gauge more accurately the useful life of antiviral medications, and the committee understands that there are ongoing efforts to do so. Beyond that, the exclusion of state and private-sector stockpiles from the federal SLEP presents a considerable barrier to further stockpiling. Recommendation 2-2: The committee recommends that the federal government’s Shelf-Life Extension Program be expanded to include other public- and private-sector entities that are stockpiling antivi- rals for use in an influenza pandemic. In the event that no solution to this issue has been found, it is pos- sible that a pandemic could coincide with availability of large state or private-sector antiviral supplies that have expired and would normally be discarded. It would be unacceptable to discard potentially viable medica- tions in a scenario of scarcity. 

34 ANTIVIRALS FOR PANDEMIC INFLUENZA Recommendation 2-3: The committee further recommends that the Department of Health and Human Services develop a process to use the knowledge acquired by the Food and Drug Administration in the operation of the Shelf-Life Extension Program to facilitate the use of properly stored, recently expired medications that exist in supplies outside the Shelf-Life Extension Program in the event these medications are needed because of a shortage. Role of the Private Sector In acknowledgment of the fact that the federal government’s abil- ity to stockpile is limited by several factors, and the fact that the private sector—specifically some major employers—are exploring ways to pro- tect some or all of their employees, DHHS has introduced the concept of “shared responsibility” and has been engaging in an ongoing dialogue with private-sector planners about their efforts (IDSA and Poretz, 2007b). The committee has learned that some private-sector employers are hesi- tant about assuming the cost of antivirals, a hesitation that is shaped by the many unknowns, the high cost of the medications, questions about shelf-life and the current exclusion of non-federal stockpiles in the SLEP, and by unease, expressed by some, at the possibility of government sei- zure of private stockpiles (Koonin, 2008). Major employers (like their counterparts in health care and in the public sector, e.g., law enforce- ment, fire departments, and emergency medical services personnel) also are uncertain about the desirability and feasibility of including workers’ families (see discussion in Chapter 4). The collective effort of many public- and private-sector entities will be needed to support the response to a pandemic. The goal of this coordina- tion will be to maximize optimal distribution of limited antiviral stock consistent with key public health strategies for pandemic containment and mitigation, including a recommendation regarding prioritization. Reliance on potential private-sector resources most likely would be dic- tated by the private sector, but the possibility of seizing private supplies for public purposes looms. Public health agency commitments not to do this might increase private-sector stockpiling (and facilitate trust in local and state public health agencies or at least a sense that all are acting in good faith). On the private-sector side, coordination may mean a commit-   As part of this program, a collaborative effort between FDA and the Department of De- fense, FDA Office of Regulatory Affairs “laboratories test product samples, and in coopera- tion with FDA’s Center for Drug Evaluation and Research, determine if the expiration date for the lot of the product can be extended and for how long” (FDA, 2007a). 

ANTIVIRAL EFFECTIVENESS, SAFETY, AND SUPPLY 35 ment to follow federal and state recommendations on antiviral dispens- ing, especially in a situation of scarcity, in which providing antivirals to low-risk employees or to those who do not perform unique and irreplace- able roles may cause public outcry and pose ethical problems. During declared states of emergency at the federal or state levels, there are legal avenues to require private-sector entities to distribute antivirals consistent with federal or state guidance. Government is sufficiently vested with emergency powers to (1) mandate that any distribution of drugs, such as antivirals, follow allocation or distribution plans; (2) change existing stan- dards of care for practitioners of medicine, nursing, pharmacy, or public health to facilitate the rapid distribution of antivirals consistent with emergency needs; or (3) take possession of available drug supplies from private-sector entities for the purpose of ensuring adherence to allocation guidelines. Any taking of private antiviral supplies must constitution- ally be compensated, although this may be determined post-emergency. Penalties for non-compliance with these potential emergency efforts may include a bevy of sanctions, such as de-licensure of medical personnel or facilities, criminal fines, and civil actions. In reality, government’s capac- ity to enforce emergency efforts will likely be compromised, necessitating collaborative efforts between public and private sectors to make possible fair and proper allocations of antivirals to assure the public’s health. Recommendation 2-4: To promote mutual trust, collaboration, and coordination, memorandums of understanding or similar agreements should be developed between public health agencies and private-sector entities in their jurisdictions. During the pre- pandemic period and in the early stages of a pandemic such collab- orations could facilitate information sharing and awareness of state and local recommendations regarding anticipated best practices in public health and standards of care in response to an influenza pandemic. (These may include prioritization schemes, guidelines for initial treatment of suspected cases, initial post-exposure pro- phylaxis, reporting of adverse events concerning antivirals, and coordination with state, tribal, and local officials as to who has been given medication.) The committee understands that in a pandemic resulting in declarations of emergency or public health emergency, federal, state, or local public health authorities have the authority to take private stockpiles of antivi- rals to meet critical societal needs, notwithstanding such agreements. Other potential areas of public–private collaboration include conduct- ing joint drills and exercises, using businesses as points-of-dispensing (Khan, 2008), and sharing knowledge and expertise, for example, 

36 ANTIVIRALS FOR PANDEMIC INFLUENZA private-sector logistical know-how that could strengthen a public health agency’s ability to move antivirals rapidly from storage to dispensing sites to people. Closing Observations The federal government and state and local public health agencies have used what is known about the severe������������������������������� ������������������������������������� influenza pandemic of 1918 as �������� ��� planning devices������������������������������������������������������������ :����������������������������������������������������������� (1) a high case fatality rate as much �������������������� ������������������������������������������������������ as 25 times that of moderate pandemics like those that occurred in 1957 and 1968; (2) high case fatality rates among the young and healthy, in addition to the very young (with somewhat lower than expected mortality among the elderly); (3) the destabilization of vital societal infrastructure, especially health care, public safety, and utilities, threatening further loss of life and social disorder; and (4) pandemic waves separated by several months. In the first wave of a severe pandemic, and in the absence of a vaccine �������������������������������� well-matched to the pandemic strain (that would take several months to develop and manufacture), ���������������������������������������������� antivirals would have the potential to affect ����������������������������������� the course of the pandemic. They could lower mortality rates across a diverse population and could help preserve critical infrastructures by pro- tecting key personnel and lessening the strain on the health care delivery system. Unfortunately�������������������������������������������������������� , confidence in the������������������������������������� ������������������������������������������������������ potential benefit������������������� s������������������ of antivirals is ����������������� tempered by the possibility of their improper use ����������������������� ������������������������������������������ (and potentially their overuse), which could perhaps����������������������������������������������� result in ������������������������������������ lessened���������������������������� benefit to the individual, and pose an enhanced risk that a resistant virus will emerge in the population. If more were known about�������������������������������� how���������������������������� antivirals����������������� would work in a pandemic������������������������������������������������������������������ , ���������������������������������������������������������������� it would be possible to make ����������������������������������� sensitive risk��������������������� –�������������������� benefit analyses of how to deploy antivirals���������������������������������������������� , and thus answer many questions that are dif- ficult, if not impossible, to answer in the current circumstances.���������� Should a limited supply of antivirals be used for treatment only, or a combination of treatment and prophylaxis? If a combination����������������������������� of the two uses is desired�� , what should be the ����������������������������������������������������� proportion������������������������������������������� of each����������������������������������� ? ��������������������������������� W�������������������������������� hat types of prophylaxis�������� should be implemented���������������������������������������������������������� , and ���������������������������������������������������� what groups should be targeted���������������������� ? Scenarios A, B, and �������������������� ���������� C answer these questions in very different ways. They represent different assumptions about the available supply of antiviral������������������� drug�������������� s, as well as about the risks associated with certain of their uses. They also factor in unknowns differently. These unknowns include, What will be the ������� effect� of community mitigation�������������������������������������������������� ������������������������������������������������������������ in terms of attack rate and mortality rate? Will the stockpiled antivirals be effective against the pandemic virus? What is the potential that������������������������������������������������������ ���������������������������������������������������������� the virus will become drug resistant����������������� and how rapidly and how extensively will resistance develop������������������������������ ? How will limited shelf-life of antivirals impact the supply at any given time? How, in the absence of 

ANTIVIRAL EFFECTIVENESS, SAFETY, AND SUPPLY 37 a timely and accurate ������������������������������������������������������ ������������������������������������������������������������������� diagnostic test, will demand for treatment antivirals be affected by the presence of other influenza������������������������������ ��������������������������������������� -like illnesses? What will be the potential for restocking antivirals, should this be affordable and ���� the� best policy objective? The greatest risk appears to be that overprescribing (e.g., to patients ������������������� who do not actually have the pandemic influenza strain) and misuse ����������� of antivirals may ������������������������������������������������� not only deplete the supply but also may promote �������� resistance, undermining ����������������������������������������������� the effectiveness of both treatment and prophy- laxis��������������������������������������������������������������������� . ������������������������������������������������������������������� Scenario����������������������������������������������������������� C poses the greatest risk in this regard. Scenario�������� ���������������� B runs the calculated risk that prophylaxis for workers providing critical societal functions (e.g., health care, utilities) does not risk over-use and misuse to the same extent. For the same reasons as the broad prophylaxis associated with scenario C, and because of its������������������������������������ ��������������������������������������� focus on��������������������������� health care and emergency services personnel, scenario B promises health care, public safety, and ��������������������������������������������������������������� social order benefits for the population as a whole.