Background

Guizhou is a mountainous province in southern China, bordered by Yunnan, Guangxi, Hunan, and Sichuan Provinces, that has a subtropical climate. The majority of the 37.5 million population live in rural villages; 11.2 million (30%) are ethnic minorities. The average population density is 22 persons/km². The province is the poorest in China, with an average gross domestic product of US $400 per capita. Zhenfeng County (population, 350,000 persons) consists of 13 townships and is located in Qianxinan Prefecture, 1 of 9 prefectures in Guizhou

In rural areas, village doctors administer OPV every 2 months. In urban areas, OPV is delivered at fixed township immunization sites at least once a month. Four doses are recommended at ages 2, 3, and 4 months and 4 years. Since 1993, Guizhou has conducted province-wide supplementary campaigns with OPV targeting all children <4 years old, regardless of prior immunization history. Although the reported coverage is >90%, actual coverage is considerably lower. A 2004 survey of 4035 children 1–3 years old in Guizhou found that the percentages of children who had received ⩾3 doses of OPV and combined diphtheria–tetanus toxoid–pertussis vaccine were 72% and 68%, respectively [3]. In 2003, reported measles incidence in Zhenfeng County was 230 cases/100,000 population, >20 times the national average

An active case-based surveillance system for poliomyelitis was established in China in 1993 in accordance with World Health Organization (WHO) guidelines to support polio eradication efforts [4]. County health department staff review records of all hospitals at the county level and above every 10 days for cases of acute flaccid paralysis (AFP) and conduct a case investigation using a standard form. Stool specimens are forwarded to provincial polio laboratories for poliovirus culture and typing [5]. Each year, 200–300 cases of AFP are reported in Guizhou, with 75%–85% having adequate specimens for culture

Poliovirus isolates are forwarded to the National Polio Laboratory in Beijing for intratypic differentiation (ITD) using molecular (polymerase chain reaction–restriction fragment–length polymorphism [PCR-RFLP]) and antigenic (ELISA) techniques [5]. PCR-RFLP has been used routinely since 1995; ELISA was introduced in 2002. A sample of isolates was retrospectively screened using ELISA, with priority given to more recent isolates, isolates from AFP case clusters or underimmunized children with residual paralysis, and isolates with discrepant PCR-RFLP results. In all, 2300 (63%) of 3629 polioviruses isolated during 1997–2004 were screened using ELISA. Nucleotide sequences of isolates with discrepant ITD results (an atypical PCR-RFLP pattern or a non–Sabin-like or double-reactive ELISA result) were sequenced at the National Polio Laboratory using methods described elsewhere [6]. Sequence data were compared with those of reference strains (GenBank). The Guizhou and National Polio laboratories are part of the WHO Global Polio Laboratory Network and are fully accredited by the WHO as national and regional reference polio laboratories, respectively

Methods

In accordance with current WHO recommendations, poliovirus isolates were categorized on the basis of the percentage difference in VP1 nucleotides from Sabin strains as follows: (1) OPV-like (<1% difference), (2) VDVP (1%–15% difference), or (3) wild (>15% difference) [2, 7]. For the purposes of the outbreak investigation, a confirmed case was defined as onset of AFP during 1 January–31 July 2004 in Guizhou, with residual paralysis and isolation of VDPV from the case patient or a contact. Regression analysis of sequence data of isolates from case patients, which was based on the known evolution rates of Sabin strains, was conducted to estimate when the initiating dose of OPV had been administered [8]. The VP1 region of all type 1 polioviruses isolated in Guizhou during January–July 2004 were sequenced to identify potential precursor strains. The entire genome of isolates from outbreak case patients and their contacts was sequenced to identify any reversion of known attenuation sites and any evidence of recombination

To determine immunization coverage, convenience surveys of children born after 1999 were conducted in affected villages in Zhenfeng and 6 adjacent counties. To verify the extent of the outbreak, records of all county- and prefecture-level hospitals in Qianxinan and Anshun Prefectures were reviewed, and a house-to-house search in affected areas was conducted to detect previously unreported cases of AFP

Information on cases of AFP and results of provincial and national polio laboratory tests since 1990 are kept in national databases maintained at the China Center for Disease Control in Beijing. We conducted a retrospective analysis of national polio surveillance data from 1997 through 2004 to identify other VDVPs isolated in China, and we collected descriptive epidemiological and laboratory information on source patients and isolates

Results

Information on outbreak case patients and their contacts is shown in table 1 [9, 10]. The index case (8233) was a 0.9-year-old boy in the village of Jiaoyang, Baiceng Township, Zhenfeng County. Jiaoyong is a remote village in the mountains that is 10 km (a 2–3-h walk) from the nearest road (figure 2). Of the 807 residents, 244 (30%) are <15 years old. On 19 May, the patient developed fever and vomiting; on 22 May, he developed bilateral weakness of the lower extremities. He was taken to the township hospital on 15 June and was reported as having AFP. Residual paralysis of the right leg was present 5 months later. Stool specimens collected from the case patient 1 month after onset tested negative for poliovirus, but a type 1 VDPV strain was isolated from stool specimens collected on 15 July from a contact, a 7.3-year-old boy (8233-c1)

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Figure 2

Location of case patients and laboratory-confirmed contacts in an outbreak of poliomyelitis caused by type 1 vaccine-derived poliovirus, Zhenfeng County, Guizhou Province, China, May–August 2004

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Table 1

Descriptive characteristics of case patients and laboratory-confirmed contacts in an outbreak of poliomyelitis caused by type 1 vaccine-derived poliovirus (VDPV), Zhenfeng County, Guizhou Province, China, May–August 2004

The second and third cases occurred in the village of Yaoshang, Wanlan Township, 40 km from Jiaoyang. Yaoshang has a resident population of 2453 and a birth rate of 17.2‰, and it lies along a paved road. The village has several coal mines that employ an additional 1000 predominantly adult migrants. The second case patient (8184) was a 3.2-year-old boy with onset of fever on 11 June and of right leg paralysis on 13 June. On 14 June, he was unable to walk and was taken to the county hospital where he was reported as having AFP. Residual paralysis was present 4 months later. A type 1 VDPV strain was isolated from stool specimens collected 5 and 6 days after the onset of AFP. The third case patient (8229) was a 1.3-year-old boy who lived 500 m from the second case patient. He developed fever, vomiting, and diarrhea on 11 July and bilateral weakness of the lower extremities on 12 July, and he was seen at a private clinic and the county hospital. On 23 July, because he had shown little improvement, he was taken to the Qianxinan Prefecture Women and Children’s Hospital and was reported as having AFP. Residual paralysis was present 3 months later. A type 1 VDPV strain was isolated from stool specimens collected 15 and 16 days after the onset of AFP. During 6–8 August, stool specimens were collected from 21 children living in Yaoshang. Type 1 VDPV strains were isolated from 3 other children who lived near the second case patient: a 3.9-year-old girl (8184-c1), her 1.4-year-old brother (8184-c3), and a 2.4-year-old girl (8184-c2) in another house

All 3 case patients and 3 of 4 laboratory-confirmed contacts had not been immunized. Investigators found that immunization services had not been delivered in Jiaoyong for several years and in Yaoshang for 6 years. Convenience surveys of children <5 years old found that 6 (100%) of 6 children in Jiaoyang and 38 (64%) of 59 children in Yaoshang had a history of never having received OPV. A coverage survey conducted in August of 619 children born after 1999 in Zhenfeng and 6 adjacent counties found that coverage with ⩾3 doses of OPV ranged from 24% to 67% and that 142 (23%) children had a history of never having received OPV. To verify immunization histories, children in the county survey were also examined for scars consistent with bacille Calmette-Guérin (BCG) immunization, which is recommended at birth for all children. Of the 529 children examined, 193 (36%) had BCG scars

None of the case patients had a history of travel outside their villages. The mother of the index case, however, went to a nearby township market several times each week. In addition, coal mined in Yaoshang is delivered to other places in Zhenfeng County, including Baiceng Township, where the index case lived

In addition to 169 cases of AFP reported in Guizhou during January–July 2004, a review of hospital records in Qianxinan and Anshun Prefectures and a search of 1257 households near Yaoshang identified an additional 10 previously unreported cases of AFP. Of the total 179 case patients identified as having AFP, 178 were discarded as nonpolio, including 138 (77%) whose stool specimens were negative. One case patient—a 0.9-year-old girl in Xingyi County, Qianxinan Prefecture—was considered to have symptoms that were clinically compatible with poliomyelitis, but stool specimens collected >2 weeks after onset tested negative. This patient had moved to a remote village 150 km from Yaoshang in November 2003 and developed fever and AFP on 29 January 2004. She had not been immunized and had residual paralysis 2 months after the onset of AFP. Immunization coverage in the village was high, with 10 (91%) of 11 children surveyed having received ⩾3 doses of OPV. Without isolates or history of contact, it is difficult to link this case patient to the outbreak

Characteristics of outbreak isolates are shown in table 2. All isolates were discrepant on ELISA (non–Sabin-like) but not PCR-RFLP and differed from reference Sabin 1 by 9–11 VP1 nt (1.0%–1.2% difference). The sequence patterns of isolates showed sequential accumulation of nucleotide substitutions consistent with sustained circulation; all isolates shared mutations in 5 sites, which suggests that they had recently branched from a single lineage. On the basis of regression analysis of sequence data and known rates of mutation of Sabin strains, it was estimated that the initiating dose of OPV had been administered in December 2003 and that the common ancestral infection of all observed isolates had occurred in March 2004, 5–6 months before the outbreak was detected. Sequencing of the complete genome of isolates from case patients did not identify evidence of recombination with other vaccine or nonpolio enteroviruses. Reversion to the wild Mahoney type was identified at 3 sites associated with attenuation of Sabin 1—at positions 2458, 2795, and 6203 nt. Sequence data from type 1 polioviruses isolated in Guizhou during January–July 2004 did not identify any precursor strains

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Table 2

Vaccine-derived poliovirus isolated from case patients with acute flaccid paralysis (AFP) and contacts in China, 1997–2004 (excludes isolates from the 2004 Guizhou outbreak shown in table 1)

A province-wide immunization response was initiated August 2004 targeting all children <5 years old with at least 2 doses of OPV. Reported coverage was >90%. Since August 2004, stool specimens from children with AFP throughout China have been negative for the outbreak strain, including specimens from 148 children with AFP in Guizhou with onset during August–December 2004

National AFP poliovirus surveillance data for 1997–2004 were retrospectively analyzed to identify other VDPVs and included the results of poliovirus culture and typing of >100,000 stool specimens collected from >50,000 case patients with AFP and their contacts throughout China. During this 8-year period, 2918 poliovirus-positive stool specimens and 3629 poliovirus isolates were identified. All isolates were screened with PCR-RFLP, and 2300 (63%) underwent ITD screening with ELISA. In all, 293 isolates had either discrepant PCR-RFLP or ELISA ITD results; of these, 270 (92%) were sequenced. In addition to the 2004 Guizhou outbreak strain, 9 other VDPV strains were identified (table 2). All 10 differed from reference Sabin strains by 9–12 VP1 nt (∼1% difference). Five were of type 1, 3 were of type 2, and 2 were of type 3. In all, 8 strains (80%) were isolated in southern China, including 3 in Guizhou, and 9 (90%) spontaneously disappeared. In 2000, a type 1 VDPV strain was isolated from 1 case patient and 2 contacts in a migrant population in Shanghai. With the exception of the 2000 Shanghai cluster and the 2004 Guizhou outbreak, the other 8 VDPV strains were sporadic isolates

Discussion

To our knowledge, this is the first outbreak of poliomyelitis in China in more than a decade and is the first caused by VDPV. The outbreak occurred in a densely populated subtropical rural area in southern China. No other cases were found, despite an intensive search, and circulation appeared to have been limited to small pockets of very low immunization coverage and population immunity. Previous VDPV strains associated with outbreaks had circulated for several years [2]. However, epidemiological and molecular data from the present investigation indicated that the outbreak strain had circulated for <1 year, which emphasizes the potential rapidity with which attenuated Sabin 1 strains can revert to the wild phenotype

Outbreaks from VDPVs have been relatively rare, with an average of 1 detected worldwide per year since 2000, and much remains unknown regarding the behavior of these viruses. Because of its large population, prolonged absence of wild poliovirus, and well-established national surveillance system, China is able to provide valuable information about the rate of emergence of VDPV strains. Although small pockets of low coverage inevitably exist, nationwide immunization coverage is high, approaching 90% [3]. Since 1997, an average of 1–2 new VDPV strains have been identified each year. Although the actual rate may be higher, because not all isolates were fully screened and the surveillance system only captures symptomatic VDPV infections, these data give an idea of the risk of emergent VDPV strains in a large population with overall high levels of immunity

The majority of VDPVs in China occurred in densely populated regions of southern China and in provinces with relatively lower levels of immunization coverage. All strains showed minimal divergence from Sabin strains and, with the exception of the outbreak in Guizhou, disappeared spontaneously. No single VDPV strain has been identified in >1 province. These data indicate that the circulation of most VDPV strains identified to date in China have been of limited duration and scope. The absence of prolonged or widespread circulation may indicate that current gaps in immunity can support the emergence of VDVP strains but not allow them to sustain transmission. Alternatively, the VDPV strains that have emerged to date may be less transmissible than wild poliovirus. The spread of a unique recombinant type 1 wild poliovirus strain from one province throughout China in <1 year indicates the potential for the circulation of a fully transmissible strain [11]

The risk of future outbreaks of VDPV infection in China, although low, may increase. Large-scale winter campaigns of immunization with OPV continue to play a key role in removing gaps in population immunity in China. These annual campaigns, however, have decreased from 30–60 million children per year during 1996–2003 to 12–16 million children per year during 2004–2005, following a shift in global eradication strategies [12]. In June 2005, a central government policy was put into effect making nationally recommended immunizations free of charge [13]. Although this marks an important step forward, any delays in fully funding the initiative could have a negative impact on immunization coverage during the transition. High routine immunization coverage, annual large-scale supplementary immunization campaigns, and sensitive and timely surveillance for wild poliovirus and VDPVs remain important government priorities to keep China polio free

In light of the risk of poliomyelitis from VDPV strains, in 2003 a WHO Expert Consultation concluded that the continued use of live OPV would compromise the goal of global eradication, and preparations have begun for the synchronized global cessation of routine OPV use shortly after confirmation of the eradication and containment of poliovirus stocks, with development of OPV stockpiles for emergency response to detected circulating wild poliovirus or VDPVs [14]. Although the risk of outbreaks of VDPV infection after global OPV cessation is estimated to be low, the actual risk is difficult to predict, because no country has ever stopped polio immunization. With an annual birth cohort of 16 million, a large susceptible population would quickly accumulate in China if immunization were stopped. The probability of an outbreak of VDPV infection in China would be much greater than that of most other countries, merely because of its large population

Many high-income countries have already switched from OPV to inactivated polio vaccine (IPV) because of safety concerns. IPV has not been seriously considered for routine use in China because of its high cost (US $2.00–$3.00/dose), compared with that of OPV (US $0.03/dose). The outbreak described in the present article and other emerging data, however, highlight the need to carefully reconsider the risks associated with OPV use when formulating future polio immunization policies for China. In addition to VDVPs, rates of vaccine-associated paralytic poliomyelitis (VAPP) in China are as high or higher than rates in other countries [15, 16]. In Shanghai and Shandong Provinces, VAPP rates were 0.5–0.8 cases/10⁶ doses of OPV administered and 3.0–7.3 cases/10⁶ first doses of OPV administered, which would translate to as many as 100–200 poliomyelitis cases/year nationwide from VAPP alone. The recent identification of long-term VDPV excretion in a boy with an underlying primary immunodeficiency disorder raises additional concerns [17]. The development of future polio immunization policies in China will require careful consideration of both current and projected risks and costs of paralytic poliomyelitis associated with VDPVs