Sabin Monovalent Oral Polio Vaccines: Review of Past Experiences and Their Potential Use after Polio Eradication

Abstract

After global eradication of polio is achieved, there will be a need for stockpiles of vaccine to combat potential outbreaks of poliomyelitis caused by (1) unforeseen release of polioviruses, (2) continued circulation of vaccine-derived strains, or (3) prolonged replication of polioviruses in immunodeficient persons. We conducted a review of the literature to document the immunogenicity and safety of monovalent Sabin vaccines, considered ideal candidates for these situations. The National Library of Medicine archives were searched for the keywords “polio,” “monovalent,” and “vaccine.” Seroconversion rates for monovalent Sabin type 1 ranged from 53% to 100% (median, 95%); for type 2, 77%–100% (median, 93%); and for type 3, 52%–100% (median, 85%). The risk of vaccine-associated poliomyelitis per million persons vaccinated ranged from .05 to 0.99 (type 1), 0–0.65 (type 2), and 1.18–8.91 (type 3). Single-dose monovalent Sabin vaccines are highly immunogenic and safe and should be considered for stockpiles of vaccine to provide an effective response to potential outbreaks of poliomyelitis in the post-eradication period.

In 1988 the World Health Assembly of the World Health Organization (WHO) resolved to eradicate poliomyelitis by 2000; since then there has been a dramatic decline in the incidence of poliomyelitis [1]. Circulation of wild poliovirus is now confined to sub-Saharan Africa and South Asia. With the expected certification of global eradication of polio in 2005 [2], planning has begun for the post-eradication period [3].

Several scientific meetings have been convened to discuss strategies for discontinuing administration of oral polio vaccine (OPV) and for biocontainment of polioviruses (both wild and vaccine-related), as well as control measures for possible poliomyelitis outbreaks in the post-eradication era [4-7]. Even after certification of polio eradication, the possibility will remain for polio outbreaks caused by (1) unforeseen release of polioviruses, either accidentally or intentionally (i.e., bioterrorism), (2) continued circulation and reversion to neurovirulence of vaccine-derived poliovirus strains, and (3) prolonged replication of polioviruses in immunodeficient persons [8-10].

One option for protecting the population in these situations is the use of homotypic monovalent vaccines, since use of trivalent oral polio vaccine (TOPV) would introduce non-outbreak-associated strains into the affected populations. We conducted a review of the literature to document the development, immunogenicity, and safety of these monovalent vaccines, which have been out of general use since the early 1960s. Because of early consensus of the superior immunogenic and safety profiles of Sabin strains over others (e.g., Koprowski and Lederle strains) and because they have been most widely used and tested, Sabin monovalent oral poliovirus vaccine (MOPV) strains are the focus of our review [11, 12].

Methods

Index Medicus for the years 1956–2000 was searched for the keywords “polio,” “monovalent,” and “vaccine.” The literature search also utilized the National Library of Medicine LOCATORplus and Centers for Disease Control and Prevention (Atlanta) electronic search software. Bibliographies from retrieved articles, books, hearings, and meeting reports were reviewed for additional references. Although large population-based studies have demonstrated “ecological effectiveness” of various monovalent Sabin strains (i.e., empirical reduction in poliomyelitis incidence occurring soon after large-scale introduction of vaccination), we limited our review of immunogenicity to seroconversion studies involving defined study cohorts. Similarly, empirical observations in many countries established the rarity of complications that are associated with monovalent Sabin strains. However, we prioritized for inclusion safety studies that used (1) an adequate surveillance system for identifying vaccine-associated paralytic poliomyelitis (VAPP), (2) a systematic process for evaluating VAPP, and (3) data that allow for assessment of the safety of MOPV, distinct from bivalent or trivalent formulations.

results

Historical Overview

Investigations of attenuated live poliovirus vaccine began in earnest in the early 1950s with studies by Dr. Hilary Koprowski demonstrating the successful immunization of volunteers [13]. Subsequent studies established that the immunogenicity of live attenuated OPV was not dependent on its neurotropic capacity [14]. Koprowski, Cox, and Sabin investigated several attenuated strains as potential vaccine candidates [15]. By 1957, the WHO had convened the Expert Committee on Poliomyelitis to define the scientific and ethical guidelines for conducting clinical trials, including the process for selection of strains and the requirements for attenuation [16]. From 1957 through 1959, the effectiveness and safety of live OPV was studied among progressively larger numbers of people, reaching in the millions by spring 1959 [17].

In 1959, the leading candidate strains, by Cox and Sabin, were submitted to the independent arbitration of Dr. Joe Melnick, a highly respected virologist. His investigation firmly established the superior safety profile of the Sabin strains and led the United States Surgeon General to officially recommend their use [12]. By the time that Sabin OPV type 3 was first licensed in the United States in 1961, Sabin vaccine strains had been administered to >50 million persons worldwide [11].

Immunogenicity

General study issues. Ideally, the effectiveness of a vaccine is evaluated through efficacy studies that measure the capacity of a vaccine to prevent disease in cohorts of nonimmune persons randomly assigned to vaccinated or nonvaccinated groups. There are no classic efficacy studies of Sabin OPV strains, for a number of reasons. A major issue was the ethical prohibition of withholding inactivated polio vaccine (IPV), which was licensed in 1955 in the United States and offered protection against a ubiquitous and potentially devastating virus.

In the absence of randomized trials, comparisons were made between self-selected vaccinated and unvaccinated groups or geographic areas after the introduction of OPV. Establishing the ideal comparison groups—those lacking prevaccination immunity—was complex, particularly in tropical countries where polioviruses circulated extensively in the early months of life. In these areas, the identification of a nonimmune cohort through serological testing of sufficient size and age representation was practically impossible. In some temperate countries, where early infection with poliovirus was less of an issue, there was the confounder of previous immunization with inactivated polio vaccine (IPV) [18]. Other confounders that had to be accounted for included (1) the natural season-to-season, year-to-year epidemic nature of poliovirus transmission, (2) the extent of circulation of other interfering enteroviruses [11], and (3) the person-to-person secondary spread, characteristic of vaccine strains, that could lessen the observed protection of Sabin strains in comparisons of vaccinated and unvaccinated persons [15].

Because of the difficulties inherent in the design of efficacy studies, the most reliable data on MOPV effectiveness are obtained from seroconversion studies evaluating the immune response produced by the vaccine. The immune response indicative of protection was most often defined as the presence of detectable antibody in a person who previously tested negative for antibody. The ideal vaccine-study participant was “triple negative,” i.e., susceptible to all poliovirus types without the confounding cross-protection that studies had shown could be induced by any one type [11, 19]. Since persons who were considered to be triple negative were relatively few in number, seroconversion studies predominantly included homotypically negative persons, who were susceptible to the administered vaccine types but not necessarily to other types.

Vaccine potency. Although data on doses were missing from a few of the studies reviewed, we found no direct relationship between vaccine potency and rates of seroconversion to MOPV types 1, 2, and 3 (figure 1). Studies involving doses of ⩾10^5.5 TCID did not show an association between dosage and greater seroconversion to types 1 and 2. We found somewhat lower seroconversion to type 3 at levels of ⩾10^5.5 TCID, although studies that used these higher doses of type 3 tended to be conducted in nontemperate countries. These findings contrast with those of studies of TOPV that have shown increased immunogenicity with vaccine containing higher relative dosages of types 1 and 3 (thereby reducing the interference of the type 2 strain) [20]. An early study by Smorodintsev et al. [21] found that intestinal titers of MOPV strain types 1–3 were unaffected by vaccine-dose reduction from 10⁵ to 10⁴ TCID but were drastically affected by further dose reduction. This study indicated a reserve on the order of 1 log associated with the Sabin recommended dose of 10⁵ TCID.

Thermostability. All monovalent Sabin strains are thermally labile [22]. Persistence of infectivity of the vaccine strain (i.e., stability) is primarily related to the use of chemical stabilizers and maintenance of the cold chain from the time of manufacture to administration of vaccine. Most of the studies that we reviewed were implemented before the discovery, in the early 1960s, of MgCl₂ as a highly effective stabilizer of poliovirus vaccine [23]; therefore, limited information is available on the use of chemical stabilizers on monovalent strains. One recent study evaluated the addition of D₂O (deuterium oxide) to MgCl₂ as a stabilizer to separate Sabin strains and found significantly improved thermostability over MgCl₂ alone [22]. There have been no trials establishing the safety and immunogenicity of D₂O use in TOPV; the need for such trials has been lessened by the advent of the vaccine-vial monitor as an extender and monitor of vaccine potency outside the cold chain [24]. It is uncertain whether monovalent strains used in the early studies would have had greater immunogenic potential had more recent chemical stabilizers been used, although one could hypothesize that the relative impact on those studies performed in tropical countries would have been greater.

Administration and scheduling. Administration of MOPV strains has occurred in various forms, volumes, and schedules. MOPV strains have predominantly been distributed in candy form in the former Soviet Union and in liquid form elsewhere. Voroshilova et al. [25] documented the immunogenicity and thermostability of monovalent Sabin strains administered in dragée candies. We identified no controlled studies comparing liquid mixtures of OPV (e.g., cherry syrup, tea, and cream-milk); however, high levels of immunogenicity were associated with most of these solutions. Similarly, we identified no controlled study demonstrating an effect of dosage volume. In a review of TOPV, Patriarca et al. [20] cited evidence that dosage volume has no consequential effect on immunogenicity.

With regard to scheduling, there are 3 factors of interest: age at first dose (namely, the issue of maternal antibody interference), interval between doses, and order of administration of vaccine types. Studies consistently documented the ability of MOPV strains to establish intestinal infections in a high percentage of neonates [26-29]. Krugman et al. [27] found that a single dose of undiluted type 1 vaccine given at birth established an intestinal infection in 8 of 10 infants. Assessing an effective immune response was complex because maternal antibodies, present in much greater relative amounts, obscured measurement of reactive antibodies in the infants.

In a randomized controlled trial comparing MOPV with TOPV and with placebo given at birth, Schoub et al. [30] found that MOPV type 1 produced greater humoral immunity (measured by geometric mean titers) than did TOPV measured throughout the first 6 months of life; this was not found for the other types. Aggregate analysis of both groups of infants (those receiving TOPV and others receiving MOPV strains) revealed that (1) moderate levels of maternal antibody (1:80–1:320) did not have an adverse impact on the seroresponse to vaccination and (2) “high levels of maternal antibodies had a marked effect on the infants' ability to mount an immune response to types 2 and 3, although they had a lesser effect for type 1” [30]. In another study of newborn infants, Lepow et al. [31] noted a blocking effect of maternal antibody at titers exceeding 1:128.

We identified no randomized controlled trials on the effect of MOPV dose interval on seroconversion. Data from the review of TOPV by Patriarca et al. [20] revealed a trend toward lower seroconversion rates when TOPV was given at 4-week intervals as opposed to 6- to 8-week intervals. A study by Pagano et al. [26] on MOPV (type 2) and non-Sabin (types 1 and 3) monovalent strains found increasing intestinal infection rates when the dose interval was increased, regardless of vaccine order. We identified no relationship between dosing interval and seroconversion, although there were few studies that used other than a 4-week interval for comparison.

Because of individual strain properties and the potential for type 2 interference, the prevalent order of administration of monovalent strains in the early vaccine trials was type 1, followed by types 3 and 2 [32, 33]. A few studies have looked at the effect of dosing sequence on immunogenicity to monovalent strains; however, almost all involved the use of either non-Sabin strains or bivalent or trivalent vaccine formulations in the dosing regimen [26, 34-36]. Only 1 study used Sabin monovalent dosing throughout (i.e., no bivalent or trivalent doses). Fornosi and Talos [37] found that when they switched the standard dosing order from 1-3-2 to 2-3-1, the seroconversion to type 2 was maintained at 100%, increased for type 1 (83% to 100%), and decreased for type 3 (75% to 69%).

Host and environmental factors. The influence of several host and environmental factors has been studied. Studies on the role of concurrent infection with other enteroviruses have generally demonstrated evidence of interference, although not uniformly [38]. Sabin and others [39-42] have shown that other concurrent enterovirus infections can interfere with the response to vaccination with MOPV strains (tables 1 and 2). It was shown in Singapore that massive administration of MOPV type 2 could significantly interfere with promulgation of an outbreak of poliomyelitis caused by a wild poliovirus type 1 [43]. In contrast, seroconversion after mass campaigns in Hungary was similar whether they were conducted in the winter or the fall, when enterovirus excretion rates were higher [44]. It is likely that the interference at least partially explains the decreased seroconversion rates to MOPV in countries where hot climates and poor sanitation are conducive to enterovirus circulation [45].

It should be noted that the studies referenced in tables 1 and 2 assessed the impact of concurrent enterovirus infection, independent of symptomatology (i.e., diarrheal illness). We did not identify any controlled studies that looked at the effect of diarrheal illness per se on seroconversion to MOPV, although recent studies have suggested that diarrheal illness may reduce the immunogenicity of TOPV [46, 47].

Assessing the impact of age on immunogenicity of monovalent Sabin strains in older children (i.e., persons >5 years of age) is difficult because of the relatively few seronegative persons in that age group during the time that MOPV strains were being evaluated [48]. The few studies evaluating older seronegative children are included in table 3. We found no evidence that age itself was a factor in the serological response to MOPV of homotypically negative older children or adults.

In assessing the impact of age on younger children aged <5 years, we identified a few studies in which age was seen to influence immunogenicity. In their evaluation of Sabin type 2 (along with CHAT type 1 and W-Fox type 3) in young infants, Pagano et al. [26] found that the magnitude of antibody response directly varied with the infant's age (independent of the presence of maternal antibody). They concluded that in spite of a high susceptibility to infection with vaccine virus in the immediate newborn period, there was a subsequent period of “relative resistance” to infection, lasting about 2–3 months [26]. Ramos Alvarez et al. [49] observed a trend toward increased seroconversion with age with types 1–3 in children aged <5 years, an association they believed could be due to the increased prevalence of other enteroviruses in the younger children. Hale et al. [43] observed a similar age-associated trend in children vaccinated with MOPV type 2 strains in Singapore.

We found no controlled studies that assessed the influence of malnutrition, concurrent illnesses, or socio-economic status on the immunogenicity to MOPV. The findings on the impact of breast-feeding were mixed. Dömök et al. [40] performed a controlled study in Uganda that evaluated the effect of breast-feeding on both intestinal infection and seroconversion after administration of Sabin type 1; no significant differences were found between infants who were breast-fed and infants who were not breast-fed. Lepow et al. [31] found that, among the cohort of infants with maternal antibodies >1:128, breast-feeding was significantly associated with decreased excretion of the vaccine virus type 1 (13 [54%] of 24 breast-fed infants vs. 48 [79%] of 61 non-breast-fed infants). Sabin also reported inhibitor effects of breast-feeding on intestinal infection with vaccine virus in neonates [50]. He found that these effects varied by virus type, with type 2 virus less susceptible than type 1 to the inhibitor effect.

Seroconversion studies. Table 3 summarizes the MOPV seroconversion studies, including information about the age of the population, vaccine formulation and administration, dosing regimens, tests and cutoffs used, study year, and location. To be included in our analysis, studies had to (1) clearly indicate that only monovalent Sabin vaccine was used, (2) determine seroconversion in persons who were homotypically seronegative to the vaccine type, and (3) assess seroconversion in at least 10 persons for any one poliovirus type. Of 16 studies that met these criteria, 12 were with Sabin monovalent types 1–3, 2 with type 1, 1 with type 2, and 1 with type 3.

The investigations were performed in 13 countries in North and South America, Europe, Africa, and Asia. Study populations ranged in size from 18v154 subjects (median, 49) for type 1; 14–135 (median, 52) for type 2; and 21–121 (median, 42) for type 3. The seroconversion rate for type 1 ranged from 53% to 100% (median, 95%); for type 2, 77%–100% (median, 93%); and for type 3, 52%–100% (median, 85%). The weighted averages for overall seroconversion rates for types 1, 2, and 3 were 90%, 93%, and 84%, respectively. Study sites were classified as to their location in temperate latitudes (north of the Tropic of Cancer, or south of the Tropic of Capricorn) or nontemperate ones. Seroconversion rates in nontemperate areas were generally lower, especially for MOPV type 3 (table 4).

Safety

Conceptual issues. The identification of an animal model for testing attenuated poliovirus strains accelerated the development of a safe vaccine. It was discovered in the 1950s that the nervous system of rhesus and cynomolgus monkeys was more susceptible to polioviruses than was the CNS in humans. Sabin further developed techniques for the propagation and selection of attenuated poliovirus strains in nonneural tissue of monkeys. The key goal was to find polioviruses with sufficient attenuation without diminishing their capacity to multiply extensively in the human gut. The attenuated properties of selected strains had to be maintained through successive manufacturing lots [17].

During 1957–1959, investigators in large-scale vaccine trials in the former Soviet Union reported the high level of effectiveness and safety associated with MOPV strains [32]. Quantifying the magnitude and consistency of adverse-event risk in different populations (for all 3 strains) was and remains challenging for several reasons. The true measure of safety of the vaccine strains is that observed after its administration to fully susceptible persons. Ideally, for measurement purposes, the persons would be triple-negative since there is a degree of cross-protection between poliovirus types [11]. Finding person who were susceptible to polio in which to assess the vaccine was especially challenging in tropical countries where infection with wild poliovirus strains occurred at a very early age [12]. Although early infection was less of an issue in temperate countries, previous immunity with inactivated polio vaccine could potentially reduce susceptibles. The need for a surveillance system to adequately investigate a rare temporal association between vaccination and serious adverse events was paramount.

Safety studies. We found several studies, performed in various countries that provided epidemiological evidence of the safety of MOPV strains, with regard to the incidence of both minor and serious adverse events (namely, VAPP). Although early investigators of MOPV strains hailed their safety, it soon became clear that there was a finite risk, albeit very small, associated with their use and that the greatest risk was associated with type 3. Noting this increased risk with type 3 in the United States, particularly among adult recipients, in 1962 the Vaccine Advisory Committee recommended the restriction of MOPV3 to preschoolers, school-age children, and adults in high-risk groups [51]. Because of the ensuing widespread use of TOPV (with a consequent decrease in the use of MOPV), there are a limited number of reports estimating the risk of VAPP associated with MOPV strains. Data from these reports were reviewed by an expert panel convened by the WHO in 1968 and were summarized in a memorandum to the WHO bulletin [52].

Our analysis of the safety of MOPV is limited to those reports that provided both numerator and denominator data (persons or doses) for evaluating risk and described a formal process for evaluating VAPP. In table 5, we summarize 4 large population-based safety studies (1 in Hungary and 3 in the United States), with information regarding the years of the studies, type and potency of Sabin vaccine given, and susceptibility of the population (if available). The risk of recipient VAPP per million persons vaccinated ranged from 0.05 to 0.99 (type 1), 0–0.65 (type 2), and 1.18–8.9 (type 3). Two studies in the United States used “per million doses” as the denominator and revealed a risk that ranged from 0.11 to 0.17 (type 1), 0.0–0.02 (type 2), and 0.4–0.8 (type 3).

In our opinion, the best data on VAPP caused by MOPV comes from Hungary, where these strains have been used the longest. Dömök described the experience in Hungary, where for >20 years MOPV had been delivered through biannual mass campaigns to children between 2 and 38 months of age [44]. These campaigns effectively eliminated indigenous transmission of wild poliovirus. Hungary had maintained an excellent surveillance system for poliomyelitis since 1966, requiring that every patient with suspected poliomyelitis be admitted to the Central Hospital for Infectious Diseases, in Budapest, for thorough clinical and laboratory evaluation. Therefore, in Hungary there existed these unique conditions: (1) a largely susceptible birth cohort; (2) absence of “confounding” protection from previous exposure to wild poliovirus strains; and (3) an excellent surveillance system.

A report by Luther Terry [51], the United States Surgeon General, summarized the data on VAPP associated with MOPV in the United States during 1961–1962. He emphasized the difficulty in establishing a causal effect between the vaccines used and poliomyelitis but also indicated that there was a high degree of evidence that Sabin MOPV type 3 posed a particularly high risk for adults. He estimated the upper limits of risk for each type by using, as the numerator, the number of VAPP cases as determined by the Vaccine Advisory Committee. Unfortunately, the denominator included a large number of nonsusceptible persons because of either previous exposure to wild poliovirus (since polio was still occurring in epidemics during that period) or vaccination with IPV.

Subsequent reports by Henderson et al. [53] and Schonberger et al. [54] summarized the United States data for the period 1962–1972, calculating risk by estimated doses administered. Limitations of these calculations include the assumptions regarding the vaccine-wastage factor (i.e., proportion of the vaccine wasted), estimated at 10%, and the proportions of age groups vaccinated. Nevertheless, comparison between the periods of 1961–1964 and 1965–1972 revealed fairly similar risks associated with each type, regardless of period, with the exception of type 3 risk, which appeared to increase by 2-fold in the latter period.

In table 6 we summarize the studies that provided risk estimates of contact VAPP cases. These results must be interpreted with great caution. Since the numerators on which these estimates are based are small, a change in classification of one case can greatly affect risk estimates. The study from Hungary provides the most reliable data since an estimate for number of susceptible contacts over an extended period is incorporated into the calculations, yielding a risk of 4.97 contact cases per million persons vaccinated [44].

Although comparisons with risk of recipient VAPP are difficult to make because of the previously cited reasons, there are important observations. First, in every study, type 3 was found to be associated with the highest risk of recipient VAPP by severalfold, followed by type 1 and type 2. Second, estimates of recipient VAPP incidence associated with type 1 and type 2 are not dissimilar to estimates in the United States associated with TOPV (0.15 per million doses) during 1980–1989 [55]. (3) Finally, the presence of a largely susceptible infant cohort in the denominator, as was the case in Hungary, was associated with considerably higher risk estimates for all types.

Discussion

From our review, we estimate that MOPV strains have been administered to at least 100 million persons worldwide. There are extensive data demonstrating high levels of single-dose immunogenicity and safety. The capacity of MOPV strains to abruptly interrupt transmission of wild poliovirus in diverse populations has been thoroughly documented [56]. On the basis of our findings, we believe that Sabin monovalent strains are excellent candidates for use against reemergent polioviruses in the post-polio eradication era.

The median rate of seroconversion in homotypically negative persons was >90% for poliovirus types 1 and 2 and 85% for poliovirus type 3 after a single dose. Seroconversion for each type was generally higher in temperate countries (table 4 and figure 1). The median seroconversion rates in nontemperate, developing countries after administration of just 1 dose of MOPV compare favorably with the median rates that Patriarca et al. [20] found in developing countries after administration of 2 doses of TOPV (78%, 88%, and 62% for types 1, 2, and 3, respectively). As with TOPV strains, the lowest rates of seroconversion to MOPV are found with vaccine virus type 3.

Unfortunately, data were not available on VAPP that occurs after administration of MOPV in tropical, developing countries. Studies of VAPP that occurs after administration of TOPV have not demonstrated a trend toward increased or decreased incidence among individuals in developing countries; estimations for first-dose recipient VAPP have been within the same general range, irrespective of level of development [57]. In Hungary, the rate of incidence of VAPP associated with for MOPV type 3 (8.9 recipient cases and 4.97 contact cases per million vaccinated) is a cause for concern, since this represents the best estimate of true risk associated with this MOPV type in a largely susceptible population in an industrialized country [44]. The first-dose recipient VAPP risk for TOPV in the United States is 1 per 750,000 persons vaccinated [13]. Whether these differences reflect differing population susceptibility profiles or vaccine delivery schemes (routine vaccine administration throughout the year in the United States vs. delivery of vaccine through mass campaigns in Hungary) is uncertain.

Reemergence of poliovirus of any type is likely to result in mass vaccination campaigns targeting hundreds of thousands, if not millions, of persons with vaccine homotypic to the circulating virus type. The VAPP cases that result from a response to a type-3 outbreak would not be inconsequential. The need for a second dose of MOPV type 3 to establish adequate levels of immunity in the population could pose an additional, although likely much smaller, risk.

Risk-benefit analyses of the use of all MOPV strains—in particular type 3—should be performed to aid in contingency planning for the post-eradication era. The use of MOPV and IPV in combination should be studied as an option to reduce VAPP cases in type-3 outbreaks. Single-dose mass administration of MOPV strains that are homotypic to a reemergent poliovirus may be sufficient in temperate, developed countries. However, a 2-dose strategy may be needed to produce high population immunity in tropical developing countries, particularly in response to reemergent type 3 strains. Another population that may be considered for a 2-dose strategy would be very young infants because of the interfering effect of maternal antibody; however, this may become less of an issue as a future cohort of previously unvaccinated women reaches childbearing age.

The risk of reintroduction of neurovirulent poliovirus strains into the population after eradication of polio will decrease with time, as remaining vaccine-derived viruses gradually disappear from the population after vaccination is discontinued. However, the ability of these reemergent strains to circulate extensively will escalate as the population becomes increasingly susceptible. The first 5–10 years after discontinuation of vaccination will be a period of especially high risk, as an age cohort known for efficient virus transmission (children <5 years old) becomes fully susceptible. Intensive surveillance for persistent vaccine-derived virus transmission will be of paramount importance.

The need for the production of large stockpiles of Sabin monovalent strains creates logistical challenges. In a WHO discussion document, Fine et al. [58] recommended that “a stockpile of at least 500 million doses of each of the three types of monovalent OPV vaccine should be maintained, and at least 100 million doses should be constantly available in 20-dose vials for immediate shipment.” They also recommended that at least 2 manufacturers maintain an ongoing capacity to produce each of the 3 monovalent strains. Although MOPVs were licensed in the 1950s and 1960s in many industrialized countries, many of these licenses have expired, and it remains unclear whether relicensing can be achieved in the current regulatory environment. One vaccine manufacturer in the United States retains a license to manufacture all 3 MOPV strains; however, the economic conditions and length of time that it would require to reestablish production capacity are uncertain.

In addition to economic considerations for vaccine manufacturers, another issue is that many industrialized countries have already adopted or intend to adopt IPV for routine vaccination against poliomyelitis. Therefore, another challenge in these countries will be to convince regulatory authorities to license a product (i.e., MOPV) that, for the foreseeable future (5–10 years), would be used primarily in developing countries. A new inactivated polio vaccine formulated with Sabin strains instead of wild strains could provide an economically viable means for producing large stocks of MOPV strains. Preliminary results of studies in Japan that have been evaluating Sabin-IPV have been promising [7].

With the certification of global eradication of polio projected for 2005 and discontinuation of vaccination expected prior to 2010, it is essential that scientific consensus on contingency plans for responding to reemergent polioviruses be reached soon [2]. Sabin monovalent strains ushered in the period of effective live-polio vaccination; it is likely they will remain the vanguard for the post-polio eradication era.

Acknowledgements

We thank Dr. Stephen Cochi, Dr. Mark Pallansch, and Dr. Patrick Zuber for their valuable assistance in reviewing the manuscript.

References