The Spread of H7N9 in China

Hong Kong’s Centre for Health Protection weekly Avian Influenza Report published on August 1 is the first weekly report in 2017 with no reported human H7N9 cases from the Republic of China. The last weekly Avian Influenza Report with no reported H7N9 cases was previously published on December 1, 2016. In the intervening 8 months more than 750 human H7N9 cases were reported from China. This is an extremely large number of human infections in a short time period.

Human cases of H7N9 were first identified and reported from China late in the 2012-2013 influenza season. Since then, there have been several waves of human H7N9 infections generally corresponding with seasonal influenza outbreaks. Although human cases of H7N9 have been sporadically reported over the past four years, the 750+ cases reported in the last eight months represent about 49% of all reported H7N9 cases. To date, all human infections have occurred in China although some individuals infected in China were identified in other countries. 

   Chart: Number of human H7N9 cases by influenza season.

As depicted in the maps below, between 13 and 16 provinces, autonomous areas, or municipalities were reporting H7N9 cases during these previous waves on a seasonal basis. This recent wave with over 750 cases, the 2016-2017 influenza season, struck China particularly hard. All provinces, autonomous areas, municipalities, or special administrative regions in China with the exception of Hainan, Heilongjiang, Qinghai, and Ningxia Hui reported at least one human H7N9 case during the past influenza season.

Map: Distribution of human H7N9 cases in China in early 2013.

Map: Distribution of human H7N9 cases in China in the 2013-2014 influenza season.

Map: Distribution of human H7N9 cases in China in the 2014-2015 influenza season.

 Map: Distribution of human H7N9 cases in China in the 2015-2016 influenza season.

Map: Distribution of human H7N9 cases in China in the 2016-2017 influenza season.

Almost all of these recent H7N9 infections are attributed to contact with infected poultry; human to human transmission is apparently very rare. There is very little public information about the distribution of low pathogenicity avian influenza (LPAI) H7N9 among domestic poultry in China. However, we can deduce from the extensive geographic distribution of human H7N9 in China this past season, that H7N9 is wide spread and perhaps endemic in poultry flocks in China.

Domestic poultry may soon be infected in adjoining countries of Myanmar, Laos, and Vietnam.  Human cases from these countries in the future would not be surprising. With the next influenza season officially starting in a few weeks, we can anticipate that additional human H7N9 cases will be reported from China in the upcoming influenza season.

The Current Status of the 2016 – 2017 H7N9 Outbreak in China as of March 1, 2017 (Is the Outbreak Just About Over?)

For the purposes of this discussion the current outbreak of H7N9 began November 1, 2016 and is still continuing. More than 460 human cases have been reported from China. Of these cases, 426 have symptom onset dates reported by the World Health Organization (WHO) for cases with onset before February 10, 2017.

Graphing the symptom onset dates for these H7N9 cases provides a count of new daily infections of H7N9. Also included in the graph are the remaining 37 cases based on their reporting date rather than symptom onset date which is not available at this time for cases reported after February 11. The graph, an epidemic curve, shows that the greatest number of H7N9 infections occurred on February 1, 2017, based on a five day moving average.

Even if The 37 cases for which symptom onset dates are not available are distributed over the 17 days following February 11, they are an insufficient number of new cases to exceed the five-day moving average which peaked above 10 cases per day on February 1, 2017. The number of human H7N9 infections in this outbreak now seems to be declining. The decline in human cases can be attributed to closing of some local poultry markets by Chinese authorities. Hopefully, the number of H7N9 infections will continue to decrease.

While there is little evidence of human-to-human transmission in this outbreak, every human H7N9 infection is a potential opportunity for the influenza virus to reassort and become transmissible between humans.

The Current Status of the 2016 – 2017 H7N9 Outbreak in China as of March 1, 2017 (Geographic Distribution)

As noted in the previous post, there have been at least 460 human cases of H7N9 reported in the current H7N9 outbreak between November 1, 2016 and February 27, 2017. Of great concern is a possibility that many of these cases are a result of human-to-human transmission. There is little publicly available information about the relationships, if any, among these hundreds of cases. To date, only four two-person clusters have been reported by the World Health Organization (WHO, January 17 and February 20) with family members comprising three of the clusters. For all four of these clusters, the WHO notes that human-to-human transmission cannot be ruled out.

One important clue to the nature of the outbreak is the geographic distribution of the reported cases. An indirect signal of human-to-human transmission can be multiple cases occurring in a localized geographic area within a short period of time. The recent WHO line listing of H7N9 cases from China (Influenza at the Human Animal Interface: Summary and Assessment, February 14, 2017), only provides the province or region for each of the reported cases. Line lists of cases provided by the Centre for Health Protection (CHP) Weekly Influenza Report provide additional geographic locational information to the prefecture level (administrative level 2) for individual cases. The Food and Agricultural Organization of the United Nations (FAO) line list of H7N9 cases occasionally provides the geographic locale of the county or administrative level 3 for some individual cases.

The most accurate locational information for individual cases is reported in local public health reports on Chinese websites. This information has been translated to English by members at FluTrackers. Sharon Sanders at FluTrackers has linked to these translated reports in the FluTrackers running list of H7N9 cases. Unfortunately, local publication of data of confirmed H7N9 cases in China are infrequent, so geographic details about individual cases beyond administrative level 2, the prefecture level, are limited to only a handful of the reported cases in this outbreak.

However, even with limited geo-locational information for individual H7N9 cases, the geographic distribution of cases can be plotted and is very informative. The map below provides a heat map of the distribution of cases in eastern China computed from the prefecture level data. Overlaid on this map are plotted locations of individual cases. The map shows the concentrations of cases in the 2016-2017 H7N9 outbreak in the provinces of Jiangsu, Zhejiang, Anhui, and Guangdong. In southern Jiangsu, hot spots include Suzhou, Wuxi, Taizhou, and Changzhou. In northern and eastern Zhejiang, the hot spots are Hangzhou, Ningbo, and Wenzhou. Hefei is the hot spot in central Anhui province and in central Guangdong, Guangzhou is the location with the most reported infections.

This map also shows that cases are widely scattered throughout many provinces during the current outbreak. The widely dispersed nature of these cases provides indirect support that human-to-human transmission is not occurring in these areas and the infections are resulting primarily from animal-to-human transmission. Even the increased number of cases in the hot spot locations does not mean that human-to-human transmission is occurring. The prefecture level cities mentioned above have very large populations most exceeding several million people. Were human-to-human transmission occurring in these areas we would expect many more reported cases.

The Current Status of the 2016 – 2017 H7N9 Outbreak in China as of March 1, 2017 (Case Count)

Since November 2016, more than 460 human cases of H7N9 have been reported or imported from China. To put this number in perspective, confirmed cases of H7N9 were first reported in March 2013, four years ago. Of all the cases of human H7N9 infections reported to date, more than one-third (about 36%), have occurred in the last four months. This raises a concern that H7N9 is not only causing outbreaks in China but could lead to epidemics and perhaps even a pandemic.

It is difficult to tabulate exactly how many H7N9 cases have occurred since November 1 of 2016, because case reporting and enumeration seem to vary among various public health reporting agencies. Media and blog reports have interpreted variation among these counts of H7N9 cases as a failure of public health officials in China to accurately track H7N9 cases, often leading to exaggerated claims of the rates of infection in China.

It is possible to arrive at a close approximation of the actual number of recent cases by using different data sources. Official counts of human cases of H7N9 are individually presented by the Food and Agricultural Organization of the United Nations (FAO), the Centre for Health Protection (CHP), and the World Health Organization (WHO). The discrepancies between these different agencies can be attributed to differential reporting periods. The FAO updates its list of human H7N9 infections every few days or whenever newly confirmed cases are reported. The CHP only updates its case information every seven days in the Weekly Influenza Report. The WHO only provides irregular updates, often only in aggregate fashion in the Disease Outbreak News. Later the WHO usually provides case details in its Influenza at the Human Animal Interface: Summary and Assessment that is only published on a monthly basis.

In order to compare these three data sets we need to have a starting point. The current outbreak in China started in November 2016. Prior to that time only a few sporadic cases were reported in the preceding weeks. Between November 1, 2016 and February 16, 2017 (the last date of FAO reported cases), the FAO has noted 437 cases of human H7N9. On November 1, 2016, the WHO count of human H7N9 cases was 800. The most recent WHO Disease Outbreak News H7N9 reports a total of 1223 confirmed H7N9 cases, indicating a total of 423 cases since November 1, 2016. The WHO case counts however only includes cases reported through February 14, 2017.

Prior to November 1, 2016, the CHP reported a total of 798 cases. Since then, the CHP has reported 461 H7N9 cases through February 27, 2017. Adjusting the FAO number of cases to include 35 cases noted by CHP with reporting dates after February 16, would bring the total FAO case count through February 27 to 472 for the period from November 1, 2016 – February 27, 2017. Adding the 43 additional cases noted by CHP (and not yet reported by WHO) to the WHO-reported count of 421 gives a total of 465 confirmed cases for the period of November 1, 2016- February 27, 2017. The variations between the adjusted counts of these three agencies is minimal, the average is 466 cases. Based on these data, the WHO count of H7N9 cases through February 27 should eventually be reported to be about 1267 cases.

The differences between the H7N9 case counts among these three agencies are primarily a function of differential reporting dates, and not the result of confusion about the number of cases by Chinese public health officials.

Another question to ask is how accurate is this count of H7N9 cases from China. In the past, China has been accused of underreporting infectious diseases to the WHO. Could there be hundreds of more human H7N9 cases that are not being reported to the WHO?

We can compare the sex ratio and median age of the 461 cases noted by the CHP in this outbreak with the sex ratio and median age reported for the previous 792 confirmed H7N9 cases prior to November 1, 2016. For the cases prior to November 1, 2016, males represent 68% of the cases, and females represent 32% of the cases. In the current outbreak, males represent 71% of the cases and females represent 29% cases. These numbers are within the range of statistical variation.

For the cases prior to November 1, 2016, males had a median age of 58 and females had a median age of 55. For the cases in the current outbreak, males have a median age of 58 and females a median age of 56. The median age for both males and females is comparable from the cases in this outbreak to all of the previous H7N9 cases reported.

Thus, there is no reason to assume that China has been underreporting H7N9 cases during this outbreak. Hypothetically, in order for China to be underreporting current H7N9 cases, the public health authorities would have to be implementing a sophisticated real-time algorithm that would allow cases to be underreported, yet still maintain the male to female ratio and the average median age for the remaining cases. Because China is reporting new H7N9 cases every few days, it does not seem possible for China to be purposefully underreporting cases of H7N9 in this outbreak. Internet claims of hundreds of unreported human H7N9 cases in China are unfounded.

What is happening with H7N9 in China?

Based on illness onset dates from January through the end of November 2016, China officially reported about 115 human cases H7N9 infection. Over the course of several days in early January 2017, China notified the World Health Organization of more than 100 additional human cases of H7N9 presumably having been infected in December 2016. It appears that almost as many people were infected in December as all of the preceding months in 2016.

The graph below shows the distribution of H7N9 cases by onset date where available and then by reporting date. The graph clearly shows the large increases in the number of infected individual reported recently. Should this increase be a cause for alarm?

Increases in human cases of avian influenza always increase the risk for sustained human to human transmission of the disease. Reviewing the minimal data that is available for the 107 recent cases reported by China, some observations can be made. About 36% of these new cases are female and 67% are male. This gender ratio is similar to the earlier cases in 2016. The age range of these cases is 23 to 91 years with a median age of 54 years old, also similar to the age distribution of earlier cases in 2016. There is no evidence from these recent cases that different age groups are being disproportionately infected.

Finally, the case fatality risk (CFR) for these recent cases is about .31. This is higher than the CFR for earlier cases in 2016 and the overall CFR for all cases since the initial outbreak in 2013. Many of the nonfatal cases are reported to have severe pneumonia, which suggests that more of these individuals may not recover.

There is nothing in the publicly available reports of these cases which would indicate the extent of human to human transmission, if any. The best indirect way to assess the potential for human to human transmission is to evaluate human clusters. In the available data, there is no information about relationships among various infected individuals, nor are onset dates available to assess chains of transmission.

 The only information we currently have available to interpret potential clusters is the geographic distribution of cases. The map below plots the geographic distribution of human cases recently reported by China compared with all of the H7N9 cases with onset dates or reported dates since January 1, 2016. About half of these newly reported cases are spread out among various provinces in eastern China and probably represented isolated sporadic infections. 

However, the remaining 50+ cases were reported from just four cities. The map below shows the four cities with 9 or more H7N9 cases reported in the January announcements, Suzhou, 21 cases, Wuxi 11 cases, and Changzhou 10 cases, all in Jiangsu Province. The fourth city is Hangzhou in Zhejiang Province with 9 cases. All of these cities are large population centers, so we will need more case details to determine if there is human to human transmission in these areas.

Information on contact tracing would be useful as well. None of the reported cases appears to be asymptomatic. Less 10 cases since the initial human H7N9 outbreak have been reported as asymptomatic. Are mild cases being overlooked?

If the number of reported H7N9 cases continues to grow dramatically over the next several weeks, it may signal a local H7N9 epidemic in China. We need to be watching H7N9 in China very closely.