Immunization with the oral poliovirus vaccine (OPV) is the cornerstone of the World Health Organization's program for the global eradication of poliomyelitis (
15,
44,
55,
56,
65). The attenuated OPV strains of the three poliovirus serotypes (Sabin 1, 2, and 3) replicate in the gut of OPV recipients and can efficiently induce type-specific humoral and mucosal immunity (
55), mimicking natural infection. However, replication of OPV in humans is frequently accompanied by genetic change of the vaccine virus, including reversion of key attenuating mutations (
5,
42), introduction of other mutations throughout the genome, and intertypic recombination among OPV strains (
7,
16). The phenotypic reversion of the OPV strains to neurovirulence is the underlying mechanism for the rare cases of vaccine-associated paralytic poliomyelitis among OPV recipients or their close contacts (
41,
54,
55). Cases of vaccine-associated paralytic poliomyelitis in immunocompetent persons are generally associated with poliovirus types 2 and 3 and very rarely with type 1 (
54). The large majority of OPV isolates from healthy individuals, the environment, or patients with vaccine-associated paralytic poliomyelitis are closely related to the original OPV strain (Sabin-like), diverging by <1.0% of nucleotide sequences encoding the major capsid protein VP1 (
8,
9,
39). The low nucleotide sequence diversities from the respective OPV strains are consistent with the short duration of most poliovirus infections (
1) and the usually restricted spread of OPV virus (
3).
It is now apparent that in areas with widening gaps in population immunity to poliovirus, especially where wild poliovirus circulation has ceased, viruses derived from OPV may circulate within a population, cause cases of paralytic poliomyelitis, and accumulate further mutations. OPV-derived polioviruses with 1 to 15% sequence divergence from the Sabin strains are now defined as vaccine-derived polioviruses (VDPVs) (
8,
9), with the extent of divergence roughly proportional to the duration of viral replication (
2,
27,
37) or circulation (
25,
30,
66) since administration of the initiating OPV dose. In 2000 to 2001, a poliomyelitis outbreak associated with circulating VDPVs (cVDPVs) occurred on the island of Hispaniola (which is divided into the Dominican Republic and Haiti), underscoring the risk of using OPV at low rates of coverage in polio-free areas (
15,
25,
28,
30,
44). Furthermore, retrospective genetic studies identified the endemic circulation of a type 2 cVDPV in Egypt from about 1983 to 1993 (
66). More recently, intensified acute flaccid paralysis and laboratory surveillance led to the identification of cVDPV outbreaks in the Philippines in 2001 (type 1) (
58,
62) and Madagascar in 2001 to 2002 (type 2) (
51).
In this report, we describe the circulation of the type 1 cVDPV in the Philippines in 2001 and the genetic and biological properties of the four cVDPV isolates. Three of the isolates were from children with acute flaccid paralysis, a common clinical manifestation of poliomyelitis. The first case was identified in March 2001 on the southern island of Mindanao in the Philippines, and two additional cases were identified in July 2001, situated about 800 km to the north on the island of Luzon (Fig.
1). The fourth isolate was from a healthy contact of one of the poliomyelitis patients in Cavite province, close to metropolitan Manila. Sequence comparisons revealed that all four cVDPVs were closely related to each other (≈99% VP1 nucleotide sequence identity), divergent from Sabin 1 (≈97% of VP1 nucleotide sequence identity), and independent of type 1 VDPVs heretofore found elsewhere. Recombinant genomes as all noncapsid sequences downstream of a common crossover site in the 2B region were derived from an as yet unidentified enterovirus. Most of the biological properties of the Philippines cVDPVs were indistinguishable from those of wild-type 1 polioviruses. Thus, the biological and genetic characteristics of the type 1 VDPVs from acute flaccid paralysis cases in the Philippines were similar to those of the cVDPVs reported from Hispaniola, Egypt, and Madagascar.
DISCUSSION
The recent cVDPV outbreak in the Philippines has important implications for the global initiative to eradicate polio. First, it demonstrates that use of OPV with suboptimal coverage rates can lead to the emergence and spread of cVDPVs even in countries where indigenous wild polioviruses have already been eradicated. Second, it again shows that use of OPV at high rates of coverage can prevent further spread of cVDPVs. Third, it reaffirms the importance of maintaining sensitive acute flaccid paralysis and poliovirus surveillance in both polio-free and polio-endemic countries. Finally, it provides additional insights into the conditions permissive for cVDPV emergence and the biological and genetic properties of the emergent viruses.
The cVDPV outbreak in the Philippines differed in key respects from earlier outbreaks reported in Egypt (
66) and Haiti in Hispaniola (
25) and the subsequent outbreak in Madagascar (
51). In the other outbreak countries, OPV coverage rates were particularly low (<50%) in the affected communities and generally low nationwide. Moreover, nearly all of the case patients in the other outbreaks were unimmunized or incompletely immunized children (
25,
51,
66). By contrast, nationwide rates of routine coverage with three doses of OPV were reported to have been approximately 80% in the Philippines since the early 1990s (
50,
62), and two of the case patients had received three doses of OPV and the third patient had received two doses (Table
1). However, gaps in population immunity probably occurred after 1997, when the mass OPV campaigns in the form of national immunization days were last conducted in the Philippines. Subnational immunization days that covered the urban areas of Manila, Cebu, and Davao (Mindanao) followed in 1998 and 1999 but did not include the three provinces with cVDPV cases (Fig.
1) (
62).
It is likely that gaps in OPV coverage developed most rapidly in the slum areas, such as those around metropolitan Manila, and these gaps were aggravated by a temporary shortage of OPV supply in 2000 to 2001. The widening immunity gap, coupled with very high population densities (especially around metropolitan Manila; Fig.
1), poor hygiene or sanitation, and tropical conditions may have established local conditions favoring cVDPV emergence. Once poliovirus circulation starts, three prior OPV doses may not be enough to protect all children from poliomyelitis, particularly in high-risk communities (
55). However, overall population immunity appears to have been sufficiently high to restrict cVDPV transmission to a minimally branched chain, in contrast to the pattern of multichain transmission seen in Egypt and Hispaniola (
25,
66). The important lesson from the Philippines outbreak is that cVDPVs can emerge even in countries with good rates of OPV coverage nationwide if immunity gaps develop in local areas at highest potential risk for poliovirus circulation.
The detection of the cVDPVs in the Philippines highlights the significant role of poliovirus surveillance in the final stages of global polio eradication. Immediately following the cVDPV outbreak in Hispaniola, intensive screening of cVDPVs was initiated by laboratories within the entire World Health Organization Global Polio Laboratory Network (
8,
9). Vaccine-related poliovirus isolates are identified by genetic methods, such as probe hybridization (
14), and also characterized for evidence of antigenic divergence from the prototype OPV strains by antigenic tests, such as intratypic differentiation-ELISA (
8,
9,
60). The likelihood of antigenic divergence increases with the duration of replication of OPV strains in the human gut (
41), and all documented cVDPV isolates have been antigenic variants of the OPV strains (
25,
51,
66). Vaccine-related isolates having altered antigenic properties are candidate VDPVs and are characterized further by genomic sequencing. In addition, the World Health Organization is promptly notified of the virologic findings in order to accelerate active surveillance and to prepare for any necessary supplementary immunization campaigns, as described here for the Philippines (
8,
9).
The Philippines cVDPV isolates, as with the other cVDPV isolates described so far (
25,
51,
66), have recombinant noncapsid sequences derived from other species C enteroviruses (
6). Since OPV contains all three serotypes of the Sabin strains, a recombinant poliovirus among heterogeneous strains readily emerges during virus replication in the gut of vaccinees. Nevertheless, recombination among the vaccine strains is known to occur frequently with serotypes 2 and 3 but rarely with type 1 (
4,
7,
13,
16,
21,
22,
38). On the other hand, circulating wild polioviruses with a block of sequence derived from Sabin 1 have been described (
34,
35). It appears most likely that the donor of the noncapsid sequences to the Philippines type 1 cVDPV isolates was nonpolio enteroviruses, as the sensitive surveillance scheme for cases of acute flaccid paralysis maintained in the Philippines has not detected any indigenous or imported wild polioviruses since 1993. Although the apparent donor of the recombinant noncapsid sequences of cVDPVs has not been identified, growing evidences indicate frequent recombination between polioviruses and species C nonpolio enteroviruses (
6,
22,
25,
28,
30,
34,
35,
51,
66) as well as between serotypes within the same nonpolio enterovirus species (
33,
36,
46,
47,
53). Further epidemiological studies of species C nonpolio enteroviruses, especially in tropical areas, are needed to understand the conditions favorable for cVDPV recombination.
Although recombination with other enteroviruses appears to be an indicator of poliovirus circulation (
25), the possible role of recombination in the phenotypic reversion of OPV is less clear. Genetic determinants of attenuation and temperature sensitivity in Sabin 1 (but not in Sabin 2 and 3) are mapped in the 3D
pol noncapsid region (
5,
12,
20,
59), so that recombination may be an efficient mechanism to replace these mutations with consensus wild enterovirus sequences. However, the major determinants of attenuation in Sabin 1 map to the 5′-NTR and capsid regions, which were not replaced by recombination in either the Philippines or Hispaniola cVDPVs.
The partially attenuated and temperature-sensitive phenotypes of the most recently identified cVDPV isolate, Luzon-01-2c, from a healthy contact child, were unexpected in view of the close sequence relationship of the contact isolate to the other three Philippines cVDPV isolates that had biological properties similar to those of wild type 1 polioviruses. The temporal and phylogenetic relationships among the Philippines cVDPV isolates suggest that isolate Luzon-01-2c was derived from a more neurovirulent and less temperature sensitive progenitor, raising the possibility that reversion of the attenuated and temperature-sensitive phenotypes of Sabin 1 is not necessarily irreversible during cVDPV evolution. Although the Luzon-01-2c isolate has the same recombinant properties as the other cVDPV isolates, it does differ from the other three isolates at some specific nucleotide and amino acid substitutions (644, 667, and 720 in the 5′-NTR, VP1-224, 2A-101, 2A-129, 2B-75, and 2C-94). Further virologic, epidemiologic, and reverse genetic studies are needed to understand the role of mutation and recombination in poliovirus evolution and cVDPV emergence.
Acknowledgments
We thank the Philippines local and regional Department of Health staff for their extensive surveillance. We also thank the virologists from the World Health Organization Polio Laboratory Network for contributing to the isolation and identification of wild polioviruses. The single-step growth experiments were performed by Ray Campagnoli and Annalet Martin (CDC). Paul Chenoweth helped prepare the map identifying the source of the cVDPV isolates. We thank Junko Wada (NIID) and Ann Turnbull (VIDRL) for excellent technical assistance.
H.S., A.U., M.A., H.Y., T.Y., and T.M. were supported, in part, by grants-in-aid for the Promotion of Polio Eradication from the Ministry of Health, Labour and Welfare, Japan. H.S. and T.Y. were supported, in part, by grants-in-aid for Development of Expanded Programme on Immunization and Accelerating Measles Control in the Polio-free Era from the Ministry of Health, Labour and Welfare, Japan. A.U. is supported by grants-in-aid from the Japan Society for the Promotion of Sciences. The Polio Regional Reference Laboratory, Australia, is supported by WHO, the Department of Health and Ageing, Canberra, and the Department of Human Services, Melbourne, Australia.