Possible meteorological influence on the Severe Acute Respiratory Syndrome (SARS) community outbreak at Amoy Gardens, Hong Kong.

Introduction

Environmental and meteorological factors are known to affect the transport, survival, and growth of many disease-causing agents (Commission on Geosciences, Environment and Resources, 2001; World Health Organization, 2004).

In the case of the severe acute respiratory syndrome (SARS) virus, Tan and co-authors (2005) found in their initial investigation that for the four cities Hong Kong, Guangzhou, Beijing, and Taiyuan in China, the optimum environmental-temperature range associated with SARS cases was 16-28[degrees]C, a range likely to be favorable to virus growth. In addition, they showed that the daily number of SARS patients was well correlated with the cumulative deviation index (CDI) of the maximum temperature observed 10 days before.

Zhang, Ye, Yang, Dong, and Zhao (2004), as well as Zhang, Yang, Ye, Xiao, and Cheng (2004), showed that for Beijing and Hong Kong, the incidence of SARS was highly correlated with air temperature and pressure. Cold-air outbreaks--that is, major temperature falls associated with the passage of cold fronts--quite likely served as a triggering mechanism. Zhang, Ye, Yang, Dong, & Zhao (2004) also developed a temperature-based High-Danger Index for SARS, and the methodology is detailed in Zhang, Yang, Ye, Xiao, and Cheng (2004). Furthermore, in statistical analysis of the climate conditions in China favorable for the occurrence of SARS, Wang and co-authors (2003) found that for the majority of the provinces, SARS was most likely to occur in autumn and spring.

Hong Kong was among the places hardest hit in the 2002-2003 global SARS pandemic, which affected more than 30 countries. Altogether, over 8,400 cases were recorded, of which some 900 resulted in deaths. In Hong Kong, out of a population of approximately 6.8 million, 1,755 cases were reported between February and June of 2003, with approximately 300 deaths. The number of fatalities in Hong Kong can be compared with 349 in the rest of China, 180 in Taiwan, 33 in Singapore, and 41 in Canada. Hong Kong was affected socially and economically as well as in terms of health (Lee, 2003). A detailed description of the epidemic in Hong Kong has been given by the SARS Expert Committee (2003).

In a "superspreading" event--that is, an event in which one case transmits to many secondary cases--some 329 people out of about 19,000 (i.e., about 1.7 percent) living in the Amoy Gardens residential estate in Hong Kong became infected with SARS between late March and early April 2003 (World Health Organization, 2003a). Forty-two of them later died. Amoy Gardens is a private residential estate for families built in 1981. It has 19 blocks, each with 33 floors. Each floor has eight apartment units arrayed around a lift lobby that forms the communal space. The size of each unit is about 48 square meters.

[FIGURE 1 OMITTED]

A number of hypotheses have been proposed with respect to the spread of SARS at Amoy Gardens. One relates to the unusual circumstances of a malfunctioning drainage system in the residential block (Block E), where the index patient visited. This malfunction caused the viral source to be aerosolized and transported into the block's lightwell (an open vertical conduit extending from the roof of the block to the ground to let air and light into the surrounding apartment units). There the aerosolized virus was carried upwards by the "chimney effect." Along the way it entered units that had windows open to the lightwell and infected some residents in these units (Department of Health, 2003; World Health Organization, 2003b). Residents infected in this way in turn infected others in Block E, as well as residents in other blocks, through person-to-person contact and contamination of the environment, and the community outbreak resulted. A second hypothesis is dissemination by roof rats (Ng, 2003), although it has been argued that this scenario was unlikely as the territorial nature of the rats would not have allowed the disease to have spread so quickly, and there were no abrupt disappearance of these rats after the SARS epidemic at Amoy Gardens had reached its peak. A third hypothesis is airborne transmission, suggested by Yu and co-authors (2004) on the basis of airflow-dynamics modeling. Roy and Milton (2004) opined that the outbreak at Amoy Gardens had at least opportunistic airborne transmission.

[FIGURE 2A OMITTED]

[FIGURE 2B OMITTED]

If airborne transmission was a pathway of the community outbreak at Amoy Gardens, meteorological factors could have contributed, and those factors are explored here. An appreciation of the role played by meteorological factors, combined with forecasts of meteorological parameters such as wind and temperature that are routinely available several days ahead, could well be useful for mitigation purposes in a recurrence, although chances of a similar outbreak are extremely remote given the unique setting at Amoy Gardens.

Our study examined the associations between peak SARS incidences at Amoy Gardens and the maximum and minimum temperatures, wind speed, wind direction, and heights of temperature inversion bases six days before, and the article discusses possible implications. Temperature inversions are layers in the atmosphere in which the air temperature increases instead of decreases with height. The heights of the temperature inversion bases are the heights to which airborne material can be transported or mixed vertically from the ground (Dobbins, 1979). They are also called mixing heights. A lag time of six days was used, as this incubation period was the mean for SARS in Hong Kong (World Health Organization, 2003a). Of course, meteorological factors, airborne transmission, and person-to-person contact might have acted together at Amoy Gardens. These pathways are not necessarily exclusive of one another.

Zhou and Yan (2003) studied the growth of SARS in Hong Kong as a whole using the Richards model. For comparison purposes, we first briefly examine the growth at Amoy Gardens using this model.

Methods

The Richards model used by Zhou and Yan (2003) is as follows:

S(t) = N/[(1 + [e.sup.-r(t-[t.sub.m])])[.sup.1/[alpha]]] (1)

where

S(t) = the cumulative number of SARS cases at day t,

N = the total number of cases, and

r = the intrinsic growth rate.

The variable [t.sub.m] is related to the point of inflection [t.sub.inflex], which is the number of days from the day of the first incidence to the time when growth reached a maximum, as follows: [t.sub.inflex] = [t.sub.m] + [r.sup.-1]ln(1/[alpha]) (Hsieh & Cheng, 2006).

[FIGURE 3A OMITTED]

The classical logistic model is given by Equation 1, with [alpha] = 1. When 1/[alpha] >1, cases are accumulating more slowly than in the logistic model and vice versa. Parameter estimation for Equation 1 was accomplished with the NLREG nonlinear regression analysis algorithm.

Data on the daily number of SARS incidences during the Amoy Gardens outbreak were extracted from the 2003 report SARS in Hong Kong: From Experience to Action, which was drawn up by the SARS Expert Committee with Sir Cyril Chantler and Professor Sian Griffiths, OBE, serving as co-chairs. This report is available at http://www.sars-expertcom.gov.hk/english/reports/reports.html.

The wind and temperature data used in our study were taken from the records of the Hong Kong Observatory The heights of the temperature inversions were taken from the 8 a.m. (local time) vertical temperature profiles obtained by balloonborne radiosondes operated at the King's Park Meteorological Station, in Hong Kong.

Results

Growth of SARS at Amoy Gardens

The first onset of SARS at Amoy Gardens was on March 14,2003. The last SARS case at Amoy Gardens was reported on April 15, 2003.

[FIGURE 3B OMITTED]

Parameter estimation with different starting values suggests that the Richards model, with 1/[alpha] =1, is appropriate for Amoy Gardens--that is, the logistic model well describes the growth of SARS at that estate. The values of r and [t.sub.m] are, respectively, 0.61 and 12.7. Thus, for Amoy Gardens, with a total number of cases, N, of 329, Equation 1 gives the following results (Figure 1):

S(t) = 329/[1 + [e.sup.-0.61(t-12.7)]] (2)

The logistic model explains 99.2 percent of the variability.

Meteorological Analysis

Figure 2a and Figure 2b show the SARS incidences at Amoy Gardens and the maximum and minimum temperatures 6 days before. The maximum temperature ranged between 16.4[degrees]C and 23.0[degrees]C and the minimum between 14.5[degrees]C and 17.9[degrees]C. These temperatures agree well with the 16[degrees]C to 28[degrees]C proposed by Tan and co-authors (2005) as optimal for SARS occurrence, and the 18[degrees]C to 22[degrees]C noted by Abdullah (2003) as permissive temperatures facilitating the transmission of the SARS virus during the Amoy Gardens outbreak.

[FIGURE 4A OMITTED]

In particular, Figure 2a shows that there was a fall of about 10[degrees]C in the maximum temperature, from 26[degrees]C to 16[degrees]C, six days before peak SARS incidences. The fall in the minimum temperature was equally substantial, about 7 [degrees]C, from 21[degrees]C to 14 [degrees]C (Figure 2b). As noted by Chang and co-authors (2005), this fall in temperature was brought about by the passage of a cold front across the south China coast on the late morning of March 18, 2003 (Figure 3a and Figure 3b). The daily maximum and minimum temperatures rose after March 20, and the number of SARS cases fell afterwards.

The passage of the cold front did not bring much change in wind direction, which mainly blew from the northeast during the days with peak SARS incidences (Figure 4a). Wind speeds, varying between 1.3 meters per second ([ms.sup.-1]) and 3.4 [ms.sup.-1] also did not strengthen after the passage of this cold front (Figure 4b). In addition, wind speeds weakened slightly from March 18 to March 21, when SARS incidences were at their peak. On the other hand, following passage of the cold front on March 18, low-level temperature inversions developed (figures 5a through 5e). On March 19, the base of this inversion, or the mixing height, was about 700 m. It fell to 550 m a day later, becoming ground based on March 21 and 22 before rising to 400 m on March 23. Cli-matologically, March in Hong Kong is the month with the most frequent occurrence of temperature inversions below 600 m (Leung, 1976).

Discussion

The cumulative growth of SARS cases at Amoy Gardens was well described by the logistic model. The values for Hong Kong as a whole, which Zhou and Yan (2003) obtained using the Richards model, were r = .09, [t.sub.m] = 6.11, and 1/[alpha] = 2.94. Thus, the growth rate of SARS at Amoy Gardens was faster than that for Hong Kong as a whole. The time to maximum growth at Amoy Gardens, 12.7 days, was also shorter than that for Hong Kong as a whole, which was 18.1 days, as calculated from [t.sub.m] + [r.sup.-1]ln(1/[alpha]) with the Richards model.

The World Health Organization (2004) has noted that many viruses do not replicate above a temperature threshold. High occurrences of other infectious diseases have been also linked to low temperatures (Kim et al., 1996), and a change in human immunity might be responsible. Thus, the fall of about 10[degrees]C in the maximum temperature and about 7[degrees]C in the minimum temperature brought about by the passage of a cold front across the south China coast in the late morning of March 18, 2003, possibly fostered the survival of the SARS virus, or rendered some of the residents more susceptible to infection, or both.

[FIGURE 4B OMITTED]

[FIGURE 5A OMITTED]

[FIGURE 5B OMITTED]

[FIGURE 5C OMITTED]

On the basis of a computational fluid-dynamics study, Yu and co-authors (2004) hypothesized that persistent northeasterly winds carried the aerosolized virus to the housing blocks downwind. A schematic diagram of the blocks at Amoy Gardens with infected people is given in Figure 6. As this figure shows, the highest rates of infections besides those in Block E were found in the downwind blocks B, C, and D. Those blocks sustained, respectively, some 13 percent, 15 percent, and 13 percent of the total number of infections, compared with about 41 percent in Block E (Department of Health, 2003). In addition, no walkways connect Block E with blocks B, C, and D, and there are also no interconnecting walkways among blocks B, C, and D, so it is reasonable to assume that one pathway by which the viral plume reached blocks B, C, and D was transport through the open air by wind. Using airflow modeling, Yu and co-authors (2004) showed that the pattern of infection in the downwind blocks was consistent with a pattern originating from a single source under the action of the prevailing northeasterly winds rather than by random person-to person contact. McKinney and co-authors (2006), in a recent review of environmental transmission at Amoy Gardens, also concluded that there was evidence to suggest that at that residential estate the virus-laden plume traveled through the air to cause human infection.

[FIGURE 5D OMITTED]

[FIGURE 5E OMITTED]

[FIGURE 6 OMITTED]

It might be ventured that had the winds strengthened after the passage of the cold front, the virus might have been carried away from Amoy Gardens more efficiently. That the winds did not pick up on this occasion was due in part to the front moving from north to south across the south China coast and Hong Kong, and the associated winds being blocked by terrain to Kong Kong's north. Since infections at Amoy Gardens also occurred in blocks other than those downwind of Block E, the airborne pathway was likely supplemented by several other possible pathways (Nicastri, Petrosillo, & Puro 2004; Tong & Liang 2004).

There was little sunshine between March 19 and 23, 2003 (Hong Kong Observatory, 2003). Thus the ground could not be heated up by solar insolation during the course of daytime to generate convection that would break up or substantially lift the temperature inversions and raise the mixing heights. In other words, the low mixing heights on the few critical dates could have contributed to the outbreak by preventing the effective dispersion of the contaminated plume of air coming out of the lightwell in Block E. Unfortunately, data on the vertical velocity and temperature of the aerosolized plume exiting the top of the light-well in Block E are not available; such data would have made possible a quantitative assessment of plume rise and relative viral concentration downwind on the basis of dispersion models as in Lighthart & Frisch (1976). Wind tunnel and etiology studies might better establish the mechanisms of transmission, but these are beyond the scope of the present investigation.

Conclusions

The SARS outbreak at Amoy Gardens in Hong Kong in March and April of 2003 had a growth pattern that is closely represented by the logistics model, with an intrinsic growth rate of 0.61 and time to maximum growth of 12.7 days.

A number of hypotheses had been suggested to explain the spread of SARS at Amoy Gardens. They include person-to person contact, dissemination by roof rats, and airborne transmission. The airborne-transmission hypothesis still needs further confirmation, but if airborne transmission was one of the pathways, then a combination of meteorological factors might have played a contributory role in the following way: Temperature falls possibly fostered the survival of the aerosolized virus, reduced immunity in some of the residents at the Amoy Gardens, or both; low mixing heights may have impeded the vertical dispersion and dilution of the aerosolized virus; and northeasterly winds may have helped to transport the virus to infect units in blocks downwind. This insight, combined with use of weather forecasts now regularly made several days ahead by meteorological services, should be useful for mitigation considerations in case of a similar occurrence, although a recurrence is very unlikely, given the unusual circumstances at Amoy Gardens.

Acknowledgements: We would like to thank the technical editor and two peer reviewers of the Journal of Enviornmental Health for very constructive comments, as well as Dr. M.C. Wu of the Hong Kong Observatory for his assistance with fitting the Richards model and Mr. Y.K. Leung and Mr. W.M. Leung of the Hong Kong Observatory for their comments on an early draft of this paper.

Corresponding Author: Professor Ignatius T.S. Yu, Department of Community and Family Medicine, 4th Floor, School of Public Health, Prince of Wales Hospital, Chinese University of Hong Kong, Shatin, New Territories, Hong Kong, China. E-mail: [e-mail].

REFERENCES

Abdullah, A.S.M. (2003). Virus pathogens suggest an autumn return. Journal of Epidemiology Community Health, 57, 770-771.

Chang, W.L., Yeung, K.H., & Leung, Y.K. (2005). Climate, severe acute respiratory syndrome (SARS) and avian flu. World Meteorological Bulletin, 54, 239-243.

Commision on Geosciences, Environment and Resources. (2001). Under the weather: climate, ecosystems, and infectious diseases. Washington D.C.: National Academy Press.

Department of Health, Government of Hong Kong Special Administration Region. (2003). Outbreak of severe acute respiratory syndrome (SARS) at Amoy Gardens, Kowloon Bay, Hong Kong: Main findings of the investigation. Retrieved April 30, 2006, from http://www.info.gov.hk/info/ap/pdf/amoy_e.pdf.

Dobbins, R.A. (1979). Atmospheric motion and air pollution: An introduction for students of engineering and science. New York: John Wiley and Sons.

Hsieh, Y.H., and Cheng, Y.S. (2006). Real-time forecast of multiphase outbreak. Emerging Infectious Diseases, 12, 122-127.

Hong Kong Observatory. (2003). Monthly weather summary, March 2003. Retrieved April 30, 2006, from http://www.weather.gov.hk/wxinfo/pastwx/mws200303.htm.

Kim, P.E., Musher, D.M., Glezen, W.P., Rodriguez-Barradas, M.C., Nahm, W.K., & Wright, C.E. (1996). Association of invasive pneumococcal disease with season, atmospheric conditions, air pollution, and the isolation of respiratory viruses. Clinical Infectious Diseases, 22(1), 100-106.

Lee, S.H. (2003). The SARS epidemic in Hong Kong. Journal of Epidemiology and Community Health, 57, 652-654.

Leung, H.C. (1976). Hong Kong upper-air climatological summaries 1961-1970 (Climatological Note No. 1). Hong Kong: Royal Observatory.

Lighthart, B., and Frisch, A.S. (1976). Estimation of viable airborne microbes downwind from a point source. Applied and Environmental Microbiology, 31, 700-704.

McKinney, Kelly R., Yu Yang Gong, Thomas G. Lewis. (2006). Environmental transmission of SARS at Amoy Gardens. Journal of Environmental Health, 68(9), 26-31.

Nicastri, E., Petrosillo, N., & Puro, V. (2004). Evidence of airborne transmission of SARS [Correspondence], New England Journal of Medicine, 351, 609-611.

Ng, S.K.C., Possible role of animal vectors in the SARS outbreak at Amoy Gardens. Lancet, 362, 570-572.

Roy, C.J., & Milton, D.K. (2004). Airborne transmission of communicable infection--The elusive pathway. New England Journal of Medicine, 350, 1710-1712.

SARS Expert Committee. (2003). SARS in Hong Kong--From experience to action: Report of the SARS Expert Committee, October 2003. Retrieved April 30, 2006, from http://www.sars-expertcom.gov.hk/english/reports/reports.html.

Tan, J., Mu, L., Huang, J., Yu, S., Chen, B., & Yin, J. (2005). An initial investigation of the association between the SARS outbreak and weather: With the view of the environmental temperature and its variation. Journal of Epidemiology and Community Health, 59,186-192.

Tong, T.R., & Liang, C. (2004). Evidence of airborne transmission of SARS [Correspondence]. New England Journal of Medicine, 351, 609-611.

Wang, Z., Cai, D., Li, S., Zheng, Y., Wang, Y, Wu, B., Li, H., & Chen, J. (2003). On season risk of the prevalence of SARS in China [In Chinese with abstract in English]. Geographical Research (Chinese Academy of Science), 22, 541-550.

World Health Organization. (2003a). Consensus document on the epidemiology of severe acute respiratory syndrome (SARS). WHO/CDS/CSR/GAR/2003.11, Retrieved April 30, 2006, from http://www.who.int/entity/csr/sars/en/WHOconsensus.pdf.

World Health Organization. (2003b). WHO Environmental Health Team reports on Amoy Gardens. Retrieved April 30, 20 from http://www.info.gov.hk/info/ap/who-amoye.pdf.

World Health Organization. (2004). Using climate to predict disease outbreaks: A review. Retrieved July 18, 2007, from http://www.who.int/globalchange/publications/oeh0401/en/index.html.

Yu, I.T.S., Li,Y., Wong, T.W, Tam, W, Chan, A.T., Lee, J.H.W, Leung, D.Y.C., & Ho, T. (2004). Evidence of airborne transmission of the severe acute respiratory syndrome virus. New England Journal of Medicine, 350, 1731-1739.

Zhang, Q., Ye, D., Yang, X., Dong, W., & Zhao, C. (2004). A high danger index for SARS incidence [In Chinese]. Chinese Journal of Public health, 20, 647-648.

Zhang, Q., Yang, X., Ye, D., Xiao F., & Cheng. Z. (2004). The meteorological characteristics and impact analysis during the period of SARS epidemic [In Chinese with English Abstract]. Journal of the Nanjing Institute of Meteorology, 19, 849-855.

Zhou, G., & Yan G. (2003). Severe acute respiratory syndrome in Asia. Emerging Infectious Diseases, 9, 1608-1610.

Although most of the information presented in the Journal refers to situations within the United States, environmental health and protection know no boundaries. The Journal periodically runs International Perspectives to ensure that issues relevant to our international constituency, representing over 60 countries worldwide, are addressed. Our goal is to raise diverse issues of interest to all our readers, irrespective of origin.

Cleo Yip, M.Phil.

Wen L. Chang, Ph.D.

K. H. Yeung

Ignatius T.S. Yu, M.P.H.