A(H5N1) and A(H7N9) have epidemiologic and virologic similarities and differences. These features have implications for detection by existing human and animal surveillance systems (Table 1).

Figure 1

Figure 1. Avian influenza A(H7N9) in humans, China, 2013–2014. Data were obtained from the World Health Organization as reported from the National Health and Family Planning Commission (http://www.who.int/influenza/human_animal_interface/influenza_h7n9/en/).

Since A(H7N9) emerged, 2 complete waves of infections have occurred; the second wave is defined as cases occurring during October 1, 2013–September 30, 2014 and affecting mostly the southeastern provinces of China (Figure 1). A third (ongoing) wave is defined as cases since October 1, 2014. A(H7N9) virus does not transmit easily between humans, and person-to-person spread has been limited to 2 or possibly 3 generations of transmission (2,4). Assuming transmission remains unchanged, geographic spread probably will occur through travel of infected humans or infected poultry. Several persons exposed to A(H7N9) virus in China traveled to Hong Kong, Taiwan, Malaysia, and Canada; became ill; and were deemed to have imported infections (2). Additional sporadic A(H7N9) infections might occur in travelers, but appropriate isolation measures should prevent further spread (2).

Figure 2

Figure 2. Initial 2-year spread of human cases and poultry outbreaks of influenza A(H5N1) in China and Southeast Asia, December 2003–2005. Data on A(H5N1) in humans were obtained from the World Health Organization...

LBMs create environments that can amplify avian influenzas viruses and increase risk of human infection (5,19,26). Large informal poultry movements between China and Southeast Asian countries also pose a risk for spread (8,27). The first confirmed human A(H5N1) case outside China occurred in December 2003 in Vietnam and foreshadowed the virus’ rapid regional spread in humans and poultry (Figure 2). Phylogeographic studies suggest that A(H5N1) virus was introduced to Vietnam from China through these poultry trade routes; A(H7N9) virus is similarly likely to be introduced in domestic poultry in Vietnam (27,28). However, because A(H7N9) preferentially infects different poultry species than A(H5N1), different poultry value chains might be implicated in this potential spread.

Wild migratory birds have contributed to spread of A(H5N1) virus along regional flyways (29,30). A(H7N9) virus was detected in a nonmigratory wild sparrow in China during spring 2013 but has not been identified in other wild bird species (13). This finding suggests that A(H7N9) infection is not widespread in wild birds, and the possible risk for regional spread by wild birds is currently low.

Figure 3

Figure 3. Severe acute respiratory infection (SARI) and influenza-like Illness (ILI) sentinel sites in Vietnam, Thailand, Laos, and Cambodia. A given location might have >1 SARI or ILI sentinel site. SARI sites in...

Vietnam, Thailand, Laos, and Cambodia each operate at least 2 national systems for influenza surveillance in humans: 1) sentinel inpatient-based severe acute respiratory infection (SARI) and 2) sentinel outpatient-based influenza-like illness (ILI). Additionally, most operate event-based or passive surveillance systems for pneumonia (Figure 3) with prespecified case definitions (Technical Appendix Table). World Health Organization member states must report all cases of A(H5N1) and A(H7N9) in humans as required by the International Health Regulations (2005); the World Organization for Animal Health mandates reporting of outbreaks of HPAI in birds by the Terrestrial Animal Health Code (31,32).

A(H7N9) and A(H5N1) viruses can cause severe disease in humans, do not transmit easily from person to person, and are primarily linked to exposure to infected poultry or contaminated environments. However, several major differences—including older age of human A(H7N9) patients, greater risk for transmission from infected poultry to humans, and lower pathogenicity of A(H7N9) virus infection in poultry—have implications for A(H7N9) virus detection. A(H7N9) will circulate in poultry without the typical HPAI-associated morbidity and mortality and can spread undetected, making A(H7N9) passive surveillance and control in poultry species challenging. Systematic, risk-based poultry surveillance is appropriate but might still miss A(H7N9) cases. As probably occurred with A(H5N1), A(H7N9) may spread through informal poultry trade between China and neighboring countries; however, the predilection of A(H7N9) for certain poultry species may favor different poultry value chains. This spread may have already occurred; although surveillance of poultry provides useful monitoring, it is unlikely to detect A(H7N9)-infected poultry until the virus becomes widespread because large sample sizes are needed to detect cases in a low-prevalence poultry population. Detection might not occur in poultry until a human case is identified, which would trigger additional poultry surveillance; notably, however, targeted poultry surveillance activities at known sites of exposure in China have had few detections despite extensive testing (20). The initial emergence of A(H5N1) in Hong Kong in 1997 was, like that of A(H7N9), associated with LBMs, but the virus has since become established in all poultry sectors, including endemicity in backyard poultry and outbreaks in wild birds (34). A(H7N9) virus might do the same, and its epidemiologic profile might change. However, the mean age of A(H5N1)-infected persons in 1997 was <10 years; no A(H5N1) cases have been reported in China in persons >65 years of age, which suggests that exposure alone cannot explain the age differences between persons with A(H5N1) and A(H7N9) infections.

Although we cannot rule out the possibility of human A(H7N9) disease in Southeast Asia, substantial A(H7N9) spread resulting in the widespread occurrence of severe infections in humans is unlikely to have occurred in Vietnam, Thailand, Cambodia, and Laos. All 4 countries have the laboratory capacity to detect A(H7N9) and well-developed hospital-based surveillance systems. Moreover, Vietnam, Thailand, and Cambodia operate passive pneumonia surveillance systems covering wide geographic areas and include public and private health care facilities. Since 2006, when Vietnam established passive surveillance, the system has been critical to detecting A(H5N1) in humans. Similarly, of 26 human A(H5N1) cases detected by CAM-EWARN in 2013, 18 sought care at private health care facilities and would not have been detected by sentinel platforms. Vietnam and Cambodia have detected A(H5N1) in humans since 2013, suggesting that severe A(H7N9) infections in humans have not occurred because either the virus has not spread to these countries or it has a lower incidence than A(H5N1). Sporadic traveler-associated A(H7N9) might be detected, but if person-to-person transmission remains limited, this mechanism should not contribute to continued spread of A(H7N9) in humans.

Although sentinel platforms in all countries may be adequate for detecting A(H7N9), their limited geographic coverage could miss cases. Efforts to expand these platforms may enhance case detection capabilities, especially in Laos, which relies solely on its sentinel surveillance systems. Therefore, robust passive reporting systems may be important to detect severe A(H7N9) cases early. Additionally, although these sentinel- and nonsentinel-based systems have performed well for A(H5N1), most detected cases occurred in children or young adults in Vietnam and Thailand or came from 1 pediatric hospital in Cambodia. Most surveillance systems cover all ages and should be able to detect illness meeting the case definition; gaps in age coverage (i.e., surveillance only at pediatric hospitals) would result in decreased sensitivity for detecting A(H7N9) because it is more likely to cause severe disease in persons >60 years of age.

Our analysis has limitations. We were unable to evaluate systems in other Asia countries; of particular concern is Myanmar, which has reported outbreaks of A(H5N1) and shares a long border with China across which it imports large quantities of poultry (3). Additionally, since July 2014, A(H7N9) has been detected in humans in Xinjiang Province in China (38), which borders 8 countries. A(H7N9) circulation in China’s western region suggests that it is probably more geographically widespread than previously realized. Additionally, our analysis focused on surveillance programs operated by governments in Vietnam, Thailand, Cambodia, and Laos. Additional surveillance may exist for avian influenza, particularly through research studies and other entities our analysis does not cover.

Since 2013, we have observed the detection of multiple novel avian influenza viruses, including A(H7N9), A(H6N1), A(H5N6), and A(H10N8) (39,40). On the basis of our assessment, we believe substantial spread of A(H7N9) virus resulting in severe infections in humans is unlikely to have occurred in Vietnam, Thailand, Cambodia, or Laos. Given the virus’ characteristics, there likely will be no obvious signal if it spreads to Southeast Asia. Well-designed, routine surveillance and astute clinicians are essential for detecting the first case beyond China. The experience with A(H5N1) shows how countries in Southeast Asia designed systems capable of detecting and responding to avian influenza in poultry and humans, but vulnerabilities remain. Growing trade networks and economic integration mean weaknesses in individual surveillance systems can leave the entire region vulnerable. Governments must be vigilant against new and reemerging disease threats by rapidly responding to suspected outbreaks in animals and humans, educating health care professionals and the public, and working with partners to enhance animal health and public health systems.