Acute flaccid paralysis incidence and Zika virus surveillance, Pacific Islands

Abstract Problem The emergence of Zika virus has challenged outbreak surveillance systems in many at-risk, low-resource countries. As the virus has been linked with Guillain–Barré syndrome, routine data on the incidence of acute flaccid paralysis (AFP) may provide a useful early warning system for the emergence of Zika virus. Approach We documented all Zika virus outbreaks and cases in 21 Pacific Islands and territories for the years 2007 to 2015. We extracted data from the Global Polio Eradication Initiative database on the reported and expected annual incidence of AFP in children younger than 15 years. Using a Poisson probability test, we tested the significance of unexpected increases in AFP in years correlating with Zika virus emergence. Data were analysed separately for each Pacific Island country and territory. Local setting In most Pacific Island countries, early warning surveillance for acute public health threats such as Zika virus is hampered by poor health infrastructure, insufficient human resources and geographical isolation. Relevant changes Only one example was found (Solomon Islands in 2015) of a significant increase in reported AFP cases correlating with Zika virus emergence. Lessons learnt We found no conclusive evidence that routinely reported AFP incidence data in children were useful for detecting emergence of Zika virus in this setting. More evidence may be needed from adult populations, who are more likely to be affected by Guillain–Barré syndrome. Reporting of AFP may be deficient in regions certified as polio-free.


Introduction
In February 2016, in response to outbreaks in several Pacific and South American countries, 1 Zika virus was declared a Public Health Emergency of International Health Concern by the World Health Organization (WHO). 2 This was based on increasing evidence that Zika virus infection may be associated with congenital malformations and autoimmune neurological diseases, including microcephaly, cranial nerve dysfunction and Guillain-Barré syndrome. 1,2 The emergence of the Zika virus has challenged basic outbreak surveillance systems in many at-risk, low-resource countries. Zika virus surveillance strategies need to be convenient, timely and cost-effective, ideally using routinely collected information. Data on the incidence of acute flaccid paralysis (AFP) in children younger than 15 years are routinely collected for polio surveillance by 177 of the 194 WHO Member States as part of the Global Polio Eradication Initiative. The most common cause of AFP is Guillain-Barré syndrome. 3 As the syndrome has been associated with Zika virus infection, increases in the incidence of AFP -routinely reported to the Global Polio Eradication Initiative -might provide a useful early warning for Zika virus outbreaks in resource-constrained settings. We analysed data from the Pacific Islands to test this hypothesis.

Local setting
The Pacific Islands, which fall within WHO's Western Pacific Region, are home to several of the world's smallest, least developed and most isolated populations. The collective population of these islands (excluding New Zealand) is approximately 11.4 million people, of whom 8.2 million reside in Papua New Guinea and the rest are dispersed over the thousands of islands and atolls that make up the other 20 Pacific Island countries and territories. Most islands' health authorities rely on simple syndromic surveillance and ad hoc event reporting by clinicians for disease outbreak detection. Their ability to enhance early warning surveillance in response to acute public health threats such as Zika virus is hampered by limited coverage and sensitivity of existing early warning surveillance; limited capacity to investigate outbreaks; geographic isolation and widely dispersed populations; poor communication infrastructure; and inadequately resourced health systems.

Approach
We documented all Zika virus outbreaks and cases in the Pacific Islands for the years 2007-2015 and tested the significance of unexpected increases in AFP in years that correlated with Zika virus emergence.
Problem The emergence of Zika virus has challenged outbreak surveillance systems in many at-risk, low-resource countries. As the virus has been linked with Guillain-Barré syndrome, routine data on the incidence of acute flaccid paralysis (AFP) may provide a useful early warning system for the emergence of Zika virus. Approach We documented all Zika virus outbreaks and cases in 21 Pacific Islands and territories for the years 2007 to 2015. We extracted data from the Global Polio Eradication Initiative database on the reported and expected annual incidence of AFP in children younger than 15 years. Using a Poisson probability test, we tested the significance of unexpected increases in AFP in years correlating with Zika virus emergence. Data were analysed separately for each Pacific Island country and territory. Local setting In most Pacific Island countries, early warning surveillance for acute public health threats such as Zika virus is hampered by poor health infrastructure, insufficient human resources and geographical isolation. Relevant changes Only one example was found (Solomon Islands in 2015) of a significant increase in reported AFP cases correlating with Zika virus emergence. Lessons learnt We found no conclusive evidence that routinely reported AFP incidence data in children were useful for detecting emergence of Zika virus in this setting. More evidence may be needed from adult populations, who are more likely to be affected by Guillain-Barré syndrome. Reporting of AFP may be deficient in regions certified as polio-free.
Lessons from the field Zika virus surveillance in the Pacific Islands Adam T Craig et al.
We obtained data on Zika virus outbreaks in the 21 Pacific Island countries and territories from published and unpublished information. For published information, we performed a literature search using the search terms "Zika" and "Zika virus" in the PubMed database; the identified papers were reviewed for relevance to the Pacific Islands. Unpublished event-relevant information was extracted from WHO's weekly Pacific Surveillance Syndromic Reports 4 and from PacNet listserv posts. 5 PacNet is the email-based outbreak notification and discussion forum of the Pacific Public Health Surveillance Network. Data extracted included the start and end dates of events, and the reported number of suspected and confirmed cases. To verify information extracted from unpublished sources we consulted staff at WHO's Division of Pacific Technical Support in Fiji. We also collected information on Zika outbreaks and cases from January to November 2016 to provide a complete and up-to-date picture of Zika activity in the study area.
AFP surveillance for polio eradication purposes requires health workers to promptly report and investigate identified cases of AFP, including the results of testing for wild polio virus. We extracted data on the expected and reported annual incidence of AFP in children younger than 15 years for Pacific Island countries and territories from the Global Polio Eradication Initiative's surveillance database. 6 Then we compared these data with documented Zika virus outbreaks to identify spacetime correlations.
We used the following Poisson probability formula to test the significance (at P ≤ 0.05) of unexpected increases in the incidence of AFP in children in the years when Zika virus emerged in each Pacific Island country or territory: Where P is the probability; X is probability mass function; x is the number of events in a specified time period; e is the mathematical constant (Euler's number, approximately 2.72); and λ is the expected number of events in the specified time period.

Zika virus cases
The first human outbreak of Zika virus was documented in the Pacific Islands in Yap State, Federated States of Micronesia in April 2007. 7 The investigators identified 185 suspected cases, of which 49 were confirmed. No further Zika cases were detected in the Pacific Islands until October 2013, when an outbreak of 383 confirmed cases occurred in French Polynesia. 8,9 Given travel pathways and close geographical and cultural ties, the outbreak in French Polynesia was suspected to have been the source of subsequent outbreaks on Easter Island (January-May 2014; 51 confirmed cases), 10 New Caledonia (January-August 2014, more than 1400 confirmed cases 11 and January-May 2015, 82 confirmed cases) 12   The significant increase in AFP cases in the Solomon Islands may be an indication of the usefulness of AFP detection for signalling the appearance of a Zika virus outbreak. Alternatively, it may reflect an increased vigilance of public health surveillance following the major tropical cyclone Raquel, which affected the Solomon Islands in July 2015, or it may just be an anomaly.

Lessons learnt
Recommendations to enhance surveillance for Zika virus in at-risk countries have included improving surveillance for Guillain-Barré syndrome via the existing surveillance systems for AFP used by polio eradication programmes. 14 Our analysis, however, did not provide sufficient evidence that analysis of AFP incidence data collected for children provide a useful surveillance strategy Populations in Pacific Island countries are small and it is likely that the capacity to conduct and ensure compliance with AFP reporting requirements varies. This may be influenced by a lack of awareness of polio surveillance and associated AFP reporting, given that the last indigenous case of polio virus in the Western Pacific Region was reported in Cambodia in 1997, and the Region has been certified as polio free since 2000. 15 Small population sizes also means that the expected incidences of AFP in individual countries and territories are very low (often < 1 case) and therefore statistical power may be lacking. More evidence is needed to determine whether the case detection of AFP is compromised in remote areas and in regions certified as polio free.
It should be noted that the Global Polio Eradication Initiative's AFP surveillance targets paediatric populations, who are less likely than adults to be affected by Guillain-Barré syndrome. 3 Data that include adult age groups (which is not currently routine practice) may provide better evidence to determine whether AFP surveillance offers a suitable strategy for Zika virus early warning in low-resource settings, such as the Pacific Islands. ■ • Routinely reported data on acute flaccid paralysis (AFP) incidence in children were insufficient for the identification of Zika virus emergence in Pacific Island countries and territories. • Data from adults, who are more affected by Guillain-Barré syndrome, may confirm whether AFP incidence is a suitable early warning surveillance strategy for emergence of Zika virus in low-resource settings. • More evidence is needed that AFP reporting requirements are being met in remote areas and in regions certified as polio free.