Skip Navigation

Transmission of the Highly Pathogenic Avian Influenza Virus H5N1 within Flocks during the 2004 Epidemic in Thailand

  1. Thanawat Tiensin1,5,
  2. Mirjam Nielen5,
  3. Hans Vernooij5,
  4. Thaweesak Songserm4,
  5. Wantanee Kalpravidh2,
  6. Sirikan Chotiprasatintara1,
  7. Arunee Chaisingh3,
  8. Surapong Wongkasemjit3,
  9. Karoon Chanachai1,
  10. Weerapong Thanapongtham1,
  11. Thinnarat Srisuvan3 and
  12. Arjan Stegeman5
  1. 1 Department of Livestock Development, Ministry of Agriculture and Cooperatives, Bangkok, Thailand
  2. 2 Food and Agriculture Organization of the United Nations, Regional Office for Asia and the Pacific, Bangkok, Thailand
  3. 3 National Institute of Animal Health, Bangkok, Thailand
  4. 4 Faculty of Veterinary Medicine, Kasetsart University, Nakhon Pathom, Thailand
  5. 5 Faculty of Veterinary Medicine, Utrecht University, Utrecht, The Netherlands
  1. Reprints or correspondence: Thanawat Tiensin Dept. of Farm Animal Health Faculty of Veterinary Medicine Utrecht University Yalelaan 7 3584 CL Utrecht The Netherlands (ttiensin{at}gmail.com, t.tiensin{at}uu.nl).
 
Next Section

Abstract

This present study is the first to quantify the transmission of avian influenza virus H5N1 within flocks during the 2004 epidemic in Thailand. It uses the flock-level mortality data to estimate the transmission-rate parameter (β) and the basic reproduction number (R0). The point estimates of β varied from 2.26/day (95% confidence interval [CI], 2.01–2.55) for a 1-day infectious period to 0.66/day (95% CI, 0.50–0.87) for a 4-day infectious period, whereas the accompanying R0 varied from 2.26 (95% CI, 2.01–2.55) to 2.64 (95% CI, 2.02–3.47). Although thepoint estimates of β of backyard chickens and fighting cocks raised together were lower than those of laying hens and broiler chickens, this difference was not statistically significant. These results will enable us to assess the control measures in simulation studies. They also indicate that, for the elimination of the virus, a critical proportion ofthe susceptible poultry population in a flock (i.e., 80% of the population) needs to be vaccinated.

Subtype H5N1 of the highly pathogenic avian influenza A (HPAI-A) virus is having a serious impact on poultry production, human health, the livelihood of farmers, wildlife conservation, and other socioeconomic factors in Asia, Africa, and Europe [1, 2]. Several measures to stop the spread of HPAI-A virus (e.g., large-scale culling, movement restrictions, vaccination, and hygienic measures) have been implemented [3–5]. However, with the exception of continental Europe and Japan, the measures have not yet resulted in the elimination of HPAI-A virus [4, 6, 7]. It has been suggested that a better understanding ofthe transmission dynamics of HPAI-A virus within flocks could help improve the effectiveness of control measures [4, 8]. In fact, quantification of the within-flock transmission could be useful for (1) determination of the time of introduction of the virus into a flock, (2) estimation of the proportion of animals requiring vaccination, and (3) assessment of the effect of intervention strategies [4, 9, 10].

Within-flock transmission can be quantified using the basic reproduction number (R0) and the transmission rate parameter (β). R0 is the expected number of secondary cases produced by a typical primary case in an entirely susceptible population [11]. R0 has a threshold value that indicates whether an infection will spread or fade out: if R0 > 1, a disease can spread; if R0 < 1, chains of transmission will inevitably fade out [11, 12]. Vaccination, one of the tools available to public health officials and veterinarians, helps reduce R0 to a value < 1, where the critical proportion of a susceptible population to be immunized is 1-1/R0 [11, 13]. β determines the average rate at which susceptible animals become infected and hence spread a disease within a population. R0 and β can be quantified from transmission experiments in which a small number of birds in a flock are inoculated with a virus. The status of all birds in the flock is then carefully monitored to record the spread of the virus [14,15].

However, outbreaks under field conditions can differ considerably from transmission experiments with a few animals under controlled conditions, in term of numbers of animals, contact structure, housing, and management. Especially when R0 > 1, which is expected in an unvaccinated population, estimates derived from transmission experiments are often imprecise [14, 15]. In the present study, the within-flock transmission of avian influenza virus H5N1 in infected chicken flocks was quantified by use of mortality data collected during the HPAI-A subtype H5N1 epidemic in Thailand. HPAI has a mortality close to 100%, and mortality data recorded in field situations are a very valuable source in its epidemiological analysis.

Here we quantified the transmission of HPAI-A virus subtype H5N1 within flocks during the 2004 epidemic in Thailand. We also investigated whether the within-flock transmission of H5N1 virus differed between flocks of backyard chickens, broilers, fighting cocks, or laying hens.

Previous SectionNext Section

Materials and Methods

Data and model assumptions. The study population included Thai chicken flocks that had tested positive for H5N1 virus by virus isolation during July–November 2004 [7]. In 139 of these flocks, mortality (number of dead birds per day) was recorded for at least 2 days before all animals were culled. These data were collected by official veterinarians of the Thai Department of Livestock Development. The 139 flocks include flocks of backyard chickens, broiler chickens, fighting cocks, and laying hens, groups that have different housing systems: backyard chickens are mixed indigenous chickens (∼20–50 birds a flock) that mingle freely in the village; broiler chickens are kept in either a closed house or an open house with netting, with large numbers of birds per housing unit, without subdivisions; laying hens are kept in battery cages in an open house with or without netting, with a large number of cages housed together; and fighting cocks are either raised with backyard chickens or sometimes kept in individual cages.

This present study was conducted under the implicit assumptions that (1) the chickens had no immunity to H5N1 virus at the time of its introduction, (2) all susceptible chickens were equally susceptible, (3) all infected chickens were equally infectious and spread the virus throughout the flock, and (4) all H5N1-infected chickens eventually died from the disease [16, 17].

Data-set construction for statistical model. The transmission data set was constructed on the basis of back-calculation, by an approach illustrated in figure 1 [8, 12, 19]. Specifically, before infection, all animals are susceptible (S), and, after infection, they become infectious (I) to other susceptible animals. Then they are removed (R) from the population, in this case because of death [12, 20]. On the basis of the literature, we assumed that chickens that had been infected started to excrete the virus 1–2 days after infection and died within 2–6 days [16, 17, 21, 22]. Estimation of β requires knowledge of both the number of susceptible chickens (S) and the number of infectious chickens (I) per day [4, 8]. Because HPAI-A is highly lethal to chickens [23, 24], we presumed that the number of dead chickens per day, in a flock, would be the number of newly infected chickens (cases) at an earlier time (C) when we back-calculated the data sets. In the present analysis, the mortality data on a flock were back-calculated into the format C(t), S(t), I(t), andR(t), for numbers of birds in each category at time t (days 0,1, 2, 3, etc.). The total number of chickens in a flock at a specific timet was also designated as N(t) = S(t) + I(t). On the basis of experimental infection studies of H5N1 virus [16, 17, 21, 22], we constructed data sets by assuming that the chickens died after an infectious period of 1, 2, 3, or 4 days (table 1).

Figure 1.
View larger version:
Figure 1.

Stages in the SIR model used to estimate avian influenza A subtype H5N1's dynamics of transmission between individual birds within a flock. A, SIR model: a bird is first susceptible (S), becomes infected (according to transmission-rate parameter [β), and stays for a time (t) in the infectious stage (I) before it dies of H5N1 (R, removed). B, Equation for the expected value for newly infected cases, used for statistical analysis [12, 18].

Figure 2.
View larger version:
Figure 2.

Characteristics of 139 chicken flocks infected with the highly pathogenic avian influenza A virus subtype H5N1 during the 2004 Thai epidemic that are used in the present study: Number and percentage of flock type (A), flock size (B), cumulative mortality (C), and average daily mortality (D).

Table 1.
View larger version:
Table 1.

Example of data-set construction of a laying hen flock, based on SIR model, at hypothetically different infectious periods of 1, 2, or 3 days.

Statistical analysis. According to this SIR assumption, the number of newly infected cases (C) per day has an expected value, E(C), that depends on β and the number of S and I animals, in the totalflock, per day. The equation for E(C) is shown in figure 1 [25, 26]. Thus, ln (β) could be estimated by a generalized linear model (GLM) using the statistical program R (version 2.2.0) [26–28]. At this stage, the back-calculated data sets (exemplified in table 1) on 139 flocks were used to estimate β. In this analysis, C was a dependent variable, and ln(S (t) I(t)/N(t)) was included as an offset variable. Flock types (i.e., backyard chickens, broiler chickens, laying hens, and fighting cocks) were added as a categorical, fixed effect, and flock was added as a random effect. The latter was needed because more observations from the same flock were included in the data set [26, 28]. We also used the negative binomial distribution, instead of a binomial or Poisson distribution, to correct for overdispersion of the outcome of the model [26, 29]. Akaike's information criterion (AIC) was used to select the best-fitting model. The model with the lowest AIC value received the most support from the data [29, 30]. Subsequently, the R0 was calculated as R0 = βT, the product of the infectious period T (days) and β (individuals per day) [4,11].

Previous SectionNext Section

Results

Characteristics of the H5N1-infected flocks. The characteristics of the 139 H5N1-infected chicken flocks used in the present study are summarized in figure 2. Most of these flocks (67%) were backyard chickens of relatively small size; whereas the laying-hen and broiler-chicken flocks were generally much larger: the flock size of the backyard chickens and fighting cocks raised together was 8–300 birds/flock, that of the laying hens was 197–7604 birds/flock, and that of the broiler chickens was 400–21,500 birds/flock. Cumulative mortality in the 139 infected flocks was 2% within 1 day of the appearance of clinical signs, with 100% mortality within 6 days and average daily mortality progressing from 1% to 36%.

Within-flock transmission. table 2 presents the outcomes of the statistical model, stratified by length of infectious period and type of chicken flock. The estimated β varied between the different infectious periods (1, 2, 3, and 4 days). Depending on the assumed infectious period for all flock types included in the model, the point estimates of β varied from 2.26/day (95% confidence interval [CI], 2.01–2.55/day) for a 1-day infectious period to 0.66/day (95% CI, 0.50–0.87/day) for a 4-day infectious period. The respective R0 values varied from 2.26 (95% CI, 2.01–2.55) to 2.64 (95% CI, 2.02–3.47) (table 2). The values of β and R0 for the different types of chicken flocks at the specified infectious periods are also shown in table 2. Although the point estimates of β and R0 for backyard chickens and fighting cocks raised together were lower than those for laying hens and broiler chickens, the difference was not statistically significant.

Table 2.
View larger version:
Table 2.

Estimates of transmission-rate parameter (β) and basic reproduction number (R0), with 95% confidence interval in parentheses, based on a generalized linear model (GLM) of data on 139 chicken flocks infected with the highly pathogenic avian influenza A virus subtype H5N1 during the 2004 Thai epidemic, with flock as a random effect and with results from the GLM with 2 groups of flock types added.

Previous SectionNext Section

Discussion

Estimates of β and R0. In the present study, we quantified the within-flock transmission of HPAI-A virus subtype H5N1 in infected chicken flocks during the 2004 epidemic in Thailand, for different types of chicken operations. The values of (resulting from this study are 0.66–2.26/day (table 2); these values are much lower than the point estimate of 33/day for a 6-day infectious period that van der Goot et al. [14] have reported for HPAI-A virus subtype H7N7. However, their estimate was based on the observations of only 2 contact-infected chickens, resulting in a CI of 1.3–∞ for R0, which includes our estimates of R0within the range of 2–5. Our value for R0 is also much lower than that used by Savill et al. [9]. In this case, comparison is difficult because their R0 estimates were modeled on the basis of data from experiments with individually challenged birds, and therefore were not quantified on the basis of either transmission data or data originating from H5N1-infected flocks in field conditions, as in our study. However, the estimate of R0 in the present study is higher than estimates for human flu pandemics (e.g., 1.89 forthe human pandemic A [H3N2] in Hong Kong [31] and 2–3 for the 1918 human pandemic A [(H1N1] in the United States [32]).

We think it very likely that the transmission of HPAI-A virus subtype H5N1 differed between the various types of flocks, because of the contact-structure differences between them, due to age, flock size, breed, management, and possibility of contact. Moreover, we expected that the differences in the density at which the chickens are housed would have an impact on transmission [4, 8, 33, 34]. However, in the present study, no significant within-flock transmission difference between flock types was found. Nevertheless, the resulting point estimates support the intuitive suggestion that within-flock transmission in backyard chickens and fighting cocks raised together is lower than that in broiler chickens and laying hens. In addition, the model with 3 groups of flocks—laying hens, broiler chickens, and backyard chickens and fighting cocks raised together—shows that the point estimates oftransmission is higher in broiler chickens than in laying hens or backyard chickens (results not shown). This would reflect the effect that contact-structure differences have on transmission. Broiler chickens are kept as 1 group, with a high possibility for mutual contact. Laying hens are kept in battery cages, with a large number of cages housed together. Backyard chickens and fighting cocks mingle freely in the village and may have lower possibility for contact than do broiler chickens and laying hens.

Bias, data validity, and back-calculation assumptions. Records of dead birds might not have been kept with the same accuracy by all poultry keepers. Moreover, record keeping might have differed depending on flock type. The number of chicken houses of large commercial poultry units, which was not known, may have caused an underestimation of the β if the disease occurred in only 1 house in a commercial unit with several houses. Data such as the total number of animals and the number of deaths per day really should be recorded per house, to facilitate epidemiologic analysis.

In our assumptions, all chickens within a flock were equally infectious for a designated period. This may differ from actual biological conditions, leading to overly precise results in the present study. Moreover, on the basis of the lethality of HPAI-A virus subtype H5N1, we assumed that all infected chickens eventually died. However, during the outbreak, culling had to be performed immediately after the infection was recognized. As a consequence, mortality data on the day of culling might not have been accurate. In our analysis, a latent-infection period (i.e., infected but not yet infectious) was not included, because, when using 1-day time interval, we assumed that an infected individual became immediately infectious [12]. Experimental studies of H5N1 have not yielded clear results on the existence or length of a latent infection period [16, 17, 21, 22]. Also, assuming an SIR model instead of an SEIR model allowed better fitting of the data (results not shown).

Application of modeling to intervention strategies. A goal of control strategies is to reduce R0 to <1. In general, R0 reduction in animal populations can be achieved in 3 ways, by reducing the (1) contact rate between animals, by isolation or separation; (2) infectiousness of infected animals, by treatment or vaccination; or (3) susceptibility of still uninfected animals, by vaccination [4, 8, 35]. When an HPAI-A epidemic in an avian species is quickly controlled, the risk of human infection is reduced [10]. Vaccination of domestic poultry against the H5N1 subtype of avian influenza has been used in several countries. Using vaccination to reduce the transmission rate could provide an alternative to mass culling, by reducing both the susceptibility of healthy flocks and the infectiousness of infected flocks. However, incomplete protection at the flock level can cause the silent spread of the virus within and between flocks. If vaccination is used but is not appropriately managed as part of a control strategy, elimination of the disease will not be achieved, and the concomitant public health risk will still be present [9, 36, 37].

In the present study, the upper limit of the R0 estimate was 5.0 (table 2), on the basis of the model with broiler chickens and laying hens used as a categorical variable (in the worst-case scenario). Therefore, the present study indicates that if at least 80% of the birds in a flock are effectively vaccinated, a major spread of the virus will not occur, on the basis of the fraction of 1-1/R0 [11]. In commercial flocks, it is feasible to apply and maintain a vaccination coverage for >80% of a total flock. However, vaccination coverage of 80% might be more difficult in backyard chickens. Another problem is that vaccine efficacy is seldom 100% [14, 16, 22]. In that case, the vaccination coverage will need to be higher than the estimated value of 1-1/R0 (i.e., >80%) but not as high as suggested by those who assume a much higher R0 [9, 14].

For other species (e.g., quails, geese, and domestic ducks), the data available are insufficient to allow transmission to be estimated. In addition, the mortality-based approach cannot be used for domestic ducks, because they show very little or late mortality [38, 39]. We suggest that transmission studies in domestic ducks be based on longitudinal serological results.

Nowadays, mathematical modeling is widely used and has become a tool for preparedness planning and for modeling of new disease outbreaks [9, 13]. The results of our analyses seemed quantitatively robust, so our β values can be used in simulation models to estimate the effectiveness of intervention measures.

In conclusion, β and R0 of HPAI-A virus subtype H5N1 were estimated by the use of mortality data collected at the flock level, from the field. This quantitative information can be used to plan a future program for control of HPAI-A.

Previous SectionNext Section

Acknowledgments

We thank the Thai Department of Livestock Development, for providing the data and for supporting this study; Monya Ekgatat, Tanawat Phansanit, Orawan Fakkham, and Mana Prasithphol, for generous help; Emelinda Lopez, Tariq Halasa, Don Klinkenberg, Michiel van Boven, and Stephen Davis, for critical comments; Linda McPhee, for editing the manuscript; and 2 reviewers, for their helpful comments.

Previous SectionNext Section

Footnotes

  • Potential conflicts of interest: none reported.

  • Financial support: Royal Government of Thailand (scholarship to T.T.); Thai Department of Livestock Development, Ministry of Agriculture and Cooperatives (support to T.T.).

  • Received December 19, 2006.
  • Accepted February 9, 2007.
Previous Section
 

References

  1. 1.↵
    Office International des Epizooties. Update on Avian Influenza in Animals (Type H5). [3 November 2006]. (Available at: http://www.who.int/csr/disease/influenza/WHO_CDS_2005_29/en/index.html.).
  2. 2.↵
    World Health Organization. Avian influenza: assessing the pandemic threat. [10 February 2006]. (Available at: http://www.who.int/csr/disease/influenza/WHO_CDS_2005_29/en/index.html.).
  3. 3.↵
    1. Capua I,
    2. Marangon S,
    3. dalla Pozza M,
    4. Terregino C,
    5. Cattoli G
    . Avian influenza in Italy 1997–2001. Avian Dis 2003;47:839-43.
    CrossRefMedlineWeb of Science
  4. 4.↵
    1. Stegeman A,
    2. Bouma A,
    3. Elbers AR,
    4. et al
    . Avian influenza A virus (H7N7) epidemic in The Netherlands in 2003: course of the epidemic and effectiveness of control measures. J Infect Dis 2004;190:2088-95.
    Abstract/FREE Full Text
  5. 5.↵
    1. Tiensin T,
    2. Chaitaweesub P,
    3. Songserm T,
    4. et al
    . Highly pathogenic avian influenza H5N1, Thailand, 2004. Emerg Infect Dis 2005;11:1664-72.
    MedlineWeb of Science
  6. 6.↵
    1. Nishiguchi A,
    2. Yamamoto T,
    3. Tsutsui T,
    4. et al
    . Control of an outbreak of highly pathogenic avian influenza, caused by the virus sub-type H5N1, in Japan in 2004. Rev Sci Tech 2005;24:933-44.
    MedlineWeb of Science
  7. 7.↵
    1. Tiensin T,
    2. Nielen M,
    3. Songserm T,
    4. et al
    . Geographic and temporal distribution of highly pathogenic avian influenza A virus (H5N1) in Thailand, 2004–05: an overview. Avian Dis 2007;51 1 Suppl:182-8.
    CrossRefMedlineWeb of Science
  8. 8.↵
    1. Stegeman A,
    2. Elbers AR,
    3. Bouma A,
    4. de Smit H,
    5. de Jong MC
    . Transmission of classical swine fever virus within herds during the 1997–1998 epidemic in The Netherlands. Prev Vet Med 1999;42:201-18.
    CrossRefMedlineWeb of Science
  9. 9.↵
    1. Savill NJ,
    2. St Rose SG,
    3. Keeling MJ,
    4. Woolhouse ME
    . Silent spread of H5N1 in vaccinated poultry. Nature 2006;442:757-757.
    CrossRefMedlineWeb of Science
  10. 10.↵
    1. Le Menach A,
    2. Vergu E,
    3. Grais RF,
    4. Smith DL,
    5. Flahault A
    . Key strategies for reducing spread of avian influenza among commercial poultry holdings: lessons for transmission to humans. Proc Biol Sci 2006;273:2467-75.
    Abstract/FREE Full Text
  11. 11.↵
    1. Anderson RM,
    2. May RM
    . Infectious diseases of humans: dynamics and control. New York: Oxford University Press; 1992.
  12. 12.↵
    1. Diekmann O,
    2. Heesterbeek JAP
    . Mathematical epidemiology of infectious diseases: model building, analysis and interpretation. Chichester: John Wiley and Sons; 2000.
  13. 13.↵
    1. Ferguson NM,
    2. Cummings DA,
    3. Cauchemez S,
    4. et al
    . Strategies for containing an emerging influenza pandemic in Southeast Asia. Nature 2005;437:209-14.
    CrossRefMedline
  14. 14.↵
    1. van der Goot JA,
    2. Koch G,
    3. de Jong MC,
    4. van Boven M
    . Quantification of the effect of vaccination on transmission of avian influenza (H7N7) in chickens. Proc Natl Acad Sci USA 2005;102:18141-6.
    Abstract/FREE Full Text
  15. 15.↵
    1. van der Goot JA,
    2. de JongMC,
    3. Koch G,
    4. Van Boven M
    . Comparison of the transmission characteristics of low and high pathogenicity avian influenza A virus (H5N2). Epidemiol Infect 2003;131:1003-13.
    Medline
  16. 16.↵
    1. Webster RG,
    2. Webby RJ,
    3. Hoffmann E,
    4. et al
    . The immunogenicity and efficacy against H5N1 challenge ofreverse genetics-derived H5N3 influenza vaccine in ducks and chickens. Virology 2006;351:303-11.
    CrossRefMedlineWeb of Science
  17. 17.↵
    1. Tian G,
    2. Zhang S,
    3. Li Y,
    4. et al
    . Protective efficacy in chickens, geese and ducks of an H5N1-inactivated vaccine developed by reverse genetics. Virology 2005;341:153-62.
    CrossRefMedlineWeb of Science
  18. 18.↵
    1. Edelstein-Keshet L
    . Mathematical models in biology. Philadelphia: Society for Industrial and Applied Mathematics; 2005.
  19. 19.↵
    1. de Jong MCM
    . Mathematical modelling in veterinary epidemiology: why model building is important. Prev Vet Med 1995;25:183-93.
    CrossRefWeb of Science
  20. 20.↵
    1. Heffernan JM,
    2. Smith RJ,
    3. Wahl LM
    . Perspectives on the basic reproductive ratio. J R Soc Interface 2005;2:281-93.
    Abstract/FREE Full Text
  21. 21.↵
    1. Lee CW,
    2. Suarez DL,
    3. Tumpey TM,
    4. et al
    . Characterization of highly pathogenic H5N1 avian influenza A viruses isolated from South Korea. J Virol 2005;79:3692-702.
    Abstract/FREE Full Text
  22. 22.↵
    1. Swayne DE,
    2. Lee CW,
    3. Spackman E
    . Inactivated North American and European H5N2 avian influenza virus vaccines protect chickens from Asian H5N1 high pathogenicity avian influenza virus. Avian Pathol 2006;35:141-6.
    CrossRefMedline
  23. 23.↵
    1. Li KS,
    2. Guan Y,
    3. Wang J,
    4. et al
    . Genesis of a highly pathogenic and potentially pandemic H5N1 influenza virus in eastern Asia. Nature 2004;430:209-13.
    CrossRefMedline
  24. 24.↵
    1. Capua I,
    2. Alexander DJ
    . Avian influenza: recent developments. Avian Pathol 2004;33:393-404.
    CrossRefMedlineWeb of Science
  25. 25.↵
    1. Becker NG
    . Analysis of infectious disease data. London: Chapman and Hall; 1989.
  26. 26.↵
    1. Dohoo I,
    2. Martin W,
    3. Stryhn H
    . Veterinary epidemiologic research. Prince Edward Island: AVC; 2003.
  27. 27.
    R Development Core Team. The R project for statistical computing. [9 March 2006]. (Available at: http://www.r-project.org).
  28. 28.↵
    1. McCullagh P,
    2. Nelder JA
    . Generalized linear models. 2nd ed. London: Chapman and Hall; 1989.
  29. 29.↵
    1. Motulsky H,
    2. Christopoulos A
    . Fitting models to biological data using linear and nonlinear regression: a practical guide to curve fitting. New York: Oxford University Press; 2004.
  30. 30.↵
    1. Akaike H
    . Information theory as an extension of the maximum likelihood principle. In: Program and abstracts of the 2nd International Symposium on Information Theory (Budapest, Hungary). Budapest, Hungary: Akademiai Kiado; 1973. p. 267-81.
  31. 31.↵
    1. Rvachev LA,
    2. Longini IM
    . A mathematical model for the global spread of influenza. Math Biosci 1985;75:3-22.
    CrossRefWeb of Science
  32. 32.↵
    1. Mills CE,
    2. Robins JM,
    3. Lipsitch M
    . Transmissibility of 1918 pandemic influenza. Nature 2004;432:904-6.
    CrossRefMedline
  33. 33.↵
    1. Keeling MJ,
    2. Woolhouse ME,
    3. May RM,
    4. Davies G,
    5. Grenfell BT
    . Modelling vaccination strategies against foot-and-mouth disease. Nature 2003;421:136-42.
    CrossRefMedline
  34. 34.↵
    1. Klinkenberg D,
    2. de Bree J,
    3. Laevens H,
    4. de Jong MC
    . Within- and between- pen transmission of classical swine fever virus: a new method to estimate the basic reproduction ratio from transmission experiments. Epidemiol Infect 2002;128:293-9.
    Medline
  35. 35.↵
    1. Stegeman A,
    2. Elbers AR,
    3. Smak J,
    4. de Jong MC
    . Quantification of the transmission of classical swine fever virus between herds during the 1997–1998 epidemic in The Netherlands. Prev Vet Med 1999;42:219-34.
    CrossRefMedlineWeb of Science
  36. 36.↵
    1. Capua I,
    2. Marangon S
    . Vaccination for avian influenza in Asia. Vaccine 2004;22:4137-8.
    CrossRefMedlineWeb of Science
  37. 37.↵
    1. Webster RG,
    2. Peiris M,
    3. Chen H,
    4. Guan Y
    . H5N1 outbreaks and enzootic influenza. Emerg Infect Dis 2006;12:3-8.
    MedlineWeb of Science
  38. 38.↵
    1. Songserm T,
    2. Jam-On R,
    3. Sae-Heng N,
    4. et al
    . Domestic ducks and H5N1 influenza epidemic, Thailand. Emerg Infect Dis 2006;12:575-581.
    MedlineWeb of Science
  39. 39.↵
    1. Sturm-Ramirez KM,
    2. Hulse-Post DJ,
    3. Govorkova EA,
    4. et al
    . Are ducks contributing to the endemicity of highly pathogenic H5N1 influenza virus in Asia? J Virol 2005;79:11269-79.
    Abstract/FREE Full Text
| Table of Contents

This Article

  1. J Infect Dis. (2007) 196 (11): 1679-1684. doi: 10.1086/522007
  1. AbstractFree
  2. » Full Text (HTML)Free
  3. Full Text (PDF)Free

Classifications

Share

  1. Email this article

Navigate This Article

Current Issue

  1. March 1, 2012 205 (5)
  1. Journal of Infectious Diseases
  1. Alert me to new issues

Most