n engl j med 348;20 www.nejm.org may 15, 2003 The new england journal of medicine 1948 antibacterial regimens for community-acquired pneumonia, and some have added a neuraminidase inhibitor to cover both influenzavirus A and influ- enzavirus B. Until we have a predictive test for the causative agent of SARS, this approach is reason- able. Supplementary oxygen should be adminis- tered if the patient has hypoxemia. The antiviral drug ribavirin has been used extensively to treat SARS, but there are no data to show that it is effec- tive. Intravenous administration was used in the pa- tients who were most ill, and oral administration (resulting in bioavailability of approximately 50 per- cent) was used in other patients. In order to use the intravenous form, a clinician in the United States must contact the CDC Emergency Operations Cen- ter (770-488-7100). Health Canada recently stated, however, that it will no longer provide access to ri- bavirin for the treatment of SARS, because of con- cern about its side effects and lack of in vitro efficacy. Some physicians have also prescribed cortico- steroids for patients with severe cases. A rationale for the use of corticosteroids derives from the patho- logical findings suggestive of cytokine dysregulation and hyperinduction of inflammatory mediators with diffuse alveolar damage. In the report by Lee et al., computed tomographic studies of the chest showed bilateral peripheral changes with ground-glass con- solidation similar to that seen in bronchiolitis ob- literans with organizing pneumonia. The latter is an inflammatory disease involving both terminal bron- chioles and alveoli that usually responds to cortico- steroids. In this time of uncertainty, we favor the use of corticosteroids only for the more ill patients. Because injectable methylprednisolone and hydro- cortisone are currently in short supply in the United States, the options are oral formulations or intrave- nous dexamethasone. SARS has created international anxiety because of its novelty, communicability, and rapid spread through jet travel and because it has caused illness in a large proportion of exposed medical and nurs- ing personnel. We simply do not know where we are on the epidemic curve. Some fear is rational, but the 4.9 percent mortality rate is in fact similar to that seen generally with community-acquired pneu- monia in the United States. Furthermore, the total number of deaths remains a small fraction of the estimated 35,000 deaths from influenza each year in the United States alone. As the epidemic unfolds, praise is due to the hun- dreds of health care workers throughout the world who come to work every day to assist patients with SARS despite some risks to their own health. Such dedication defines the best traditions of our pro- fession. From the Department of Internal Medicine, Virginia Common- wealth University, Richmond. SARS-Associated Coronavirus Kathryn V. Holmes, Ph.D. The discovery that a novel coronavirus is the prob- able cause of the newly recognized severe acute res- piratory syndrome (SARS), reported by Ksiazek et al. (pages 1953–1966), Drosten et al. (pages 1967– Management of Suspected SARS. Isolate the patient Place patient in private room (with negative pres- sure, if possible) Wear gloves, gown, masks, eye protection Wash hands carefully after removing gloves Limit number of health care workers caring for patient Limit number of visitors Perform diagnostic studies Obtain specimens to rule out causes of atypical pneumonia Obtain specimens for SARS testing (see CDC Web page, http://www.cdc.gov/ncidod/sars/ specimens.htm) Consider computed tomography of chest Provide treatment Supplementary oxygen for hypoxemia Antibacterial agents for community-acquired pneumonia Consider neuraminidase inhibitor for treatment of influenza Ribavirin (oral formulation: 1.2 g every 8 hr; commercially available, intravenous form: 8 mg/kg of body weight every 8 hr) (available through the CDC) Consider corticosteroids Notify public health department The New England Journal of Medicine Downloaded from nejm.org on May 23, 2014. For personal use only. No other uses without permission. Copyright © 2003 Massachusetts Medical Society. All rights reserved. n engl j med 348;20 www.nejm.org may 15, 2003 perspective 1949 1976), and Peiris et al.1 provides a dramatic exam- ple of an emerging coronavirus disease in humans, described by Poutanen et al. (pages 1995–2005), Tsang et al. (pages 1977–1985), and Lee et al. (pag- es 1986–1994). Although human coronaviruses cause up to 30 percent of colds, they rarely cause lower respiratory tract disease. In contrast, coro- naviruses cause devastating epizootics of respira- tory or enteric disease in livestock and poultry. Most coronaviruses cause disease in only one host species. All known coronaviruses are found in three serologically unrelated groups. The Figure shows the structure of the virion. The message- sense RNA genome and the viral nucleocapsid phosphoprotein form a helical nucleocapsid. A co- rona of large, distinctive spikes in the envelope makes possible the identification of coronavirus- es by electron microscopy. The spikes, oligomers of the spike(S) glycoprotein, bind to receptors on host cells and fuse the viral envelope with host cell membranes. Coronaviruses in group 2 also have a hemagglutinin–acetylesterase (HE) glycoprotein that binds to sugar moieties on cell membranes. Curiously, the gene for HE was apparently intro- duced into an ancestral coronavirus genome by re- combination with the messenger RNA encoding HE of influenza C. The unique RNA-dependent RNA polymerase of coronaviruses often switches tem- plate strands during replication, causing RNA re- combination when a cell is infected with several coronaviruses. This error-prone polymerase also generates point mutations and large deletions or insertions of foreign RNA into the viral genome. The SARS-associated coronavirus could have arisen as a mutant of a human coronavirus that ac- quired new virulence factors, as a mutant of an an- Structure of the Coronavirus Virion. Spike glycoprotein Small envelope glycoprotein Membrane glycoprotein Hemagglutinin–acetylesterase glycoprotein Nucleocapsid phosphoprotein RNA The New England Journal of Medicine Downloaded from nejm.org on May 23, 2014. For personal use only. No other uses without permission. Copyright © 2003 Massachusetts Medical Society. All rights reserved. n engl j med 348;20 www.nejm.org may 15, 2003 The new england journal of medicine 1950 imal coronavirus that can infect human cells, or as a recombinant of two human coronaviruses or a hu- man coronavirus and an animal coronavirus. Anti- bodies to the SARS-associated coronavirus were found in serum samples obtained from patients with SARS during convalescence but not in human serum samples banked before the SARS outbreak, suggesting that the SARS-associated coronavirus is new to the human population. The nucleotide se- quence of the SARS-associated coronavirus genome (http://www.bcgsc.ca/bioinfo/SARS; http://www. cdc.gov/ncidod/sars/sequence.htm) differs sub- stantially from sequences of all known coronavi- ruses. Thus, the SARS-associated coronavirus is neither a mutant of any known coronavirus nor a recombi- nant of known coronaviruses. It is a previously un- known coronavirus, probably from a nonhuman host, that somehow acquired the ability to infect humans. Serologic tests of wild and domestic ani- mals and birds in the region where the outbreak first appeared may identify the usual host. Comparison of isolates of the SARS-associated coronavirus from infected patients and from the natural host may re- veal how the virus jumped to humans. In jumping to humans, did the SARS-associated coronavirus lose the ability to infect its original host? If there is no animal reservoir, there will be a better chance of eliminating the virus from humans. The host range, tissue tropism, and virulence of animal coronaviruses can be changed by mutations in the S gene. The sequence of the S gene in the SARS-associated coronavirus may suggest how S glycoprotein affects the pathogenesis of SARS. The SARS-associated coronavirus genome sequence shows that it does not contain a gene encoding HE or large genes derived from another virus or host cell. It is an amazing feat that the SARS-associated coronavirus genome has been completely se- quenced so quickly. The surprising discovery that the virus can be readily isolated in a monkey-kidney cell line was the key to the rapid molecular charac- terization of this novel coronavirus and the devel- opment of diagnostic tests for SARS. SARS-associ- ated coronavirus has recently been proved to be the cause of SARS. Inoculation of monkeys with SARS- associated coronavirus from cell cultures caused lower respiratory tract disease, fulfilling Koch’s postulate. Both viral and host factors affect the virulence of coronavirus diseases in animals. The disease is usu- ally most severe in neonates. The signs of infection in immunosuppressed animals may differ from those in immunocompetent animals; immunosup- pressed animals may also shed virus for prolonged periods and accumulate and possibly spread mutant viruses. The detection of SARS-associated corona- virus in fecal and serum samples from patients, as well as in respiratory specimens, suggests that this virus, like many animal coronaviruses, may be spread both by fecal contamination and by respira- tory droplets. Host genes that affect the viral recep- tor, viral production, and immune responses to in- fection can determine the outcome of coronavirus infections, making certain species or strains of an- imals highly susceptible to lethal infection. For ex- ample, coronaviruses from domestic cats almost always cause death in cheetahs. Coinfection with other viruses, parasites, or bacteria exacerbates some animal coronavirus diseases. The deaths of 3 to 4 percent of patients with SARS may result from host factors that exacerbate the disease. Although there are no approved drugs with prov- en efficacy against coronaviruses, there are poten- tial targets for the development of new drugs. Pro- tease inhibitors could prevent processing of the RNA polymerase or cleavage of the viral S glycopro- tein. Inhibitors of coronavirus acetylesterase activity might limit viral replication, as neuraminidase in- hibitors inhibit the replication of influenzaviruses A and B. Inhibitors of membrane fusion might block viral entry, as do several new drugs against the human immunodeficiency virus. Antibodies against the viral S glycoprotein or the unidentified receptor for the SARS-associated coronavirus might also block entry of the virus. Vaccines are available for some animal corona- viruses. Vaccination with live, attenuated virus is ef- fective against porcine epidemic diarrhea virus and avian infectious bronchitis virus. However, recom- bination of genomes of vaccine strains with wild- type coronaviruses is a potential risk associated with using live, attenuated coronavirus vaccines in hu- mans. Killed or subunit vaccines containing the spike glycoprotein, perhaps with other viral pro- teins, might prevent lower respiratory tract disease in humans. However, some vaccines against feline coronaviruses actually enhanced disease when vac- cinated animals were exposed to wild-type virus, and antibody enhancement of disease is a potential risk of SARS vaccines in humans. It is possible that the current outbreak may be controlled and the virus eliminated by quarantine alone. Nevertheless, it is prudent to develop safe, effective drugs and vaccines The New England Journal of Medicine Downloaded from nejm.org on May 23, 2014. For personal use only. No other uses without permission. Copyright © 2003 Massachusetts Medical Society. All rights reserved. n engl j med 348;20 www.nejm.org may 15, 2003 perspective 1951 against the Urbani SARS-associated coronavirus as quickly as possible, in case the outbreak cannot be contained. The development of drugs and vac- cines for SARS will also provide new strategies for the prevention and treatment of other coronavirus diseases of animals and humans. From the University of Colorado Health Sciences Center, Denver. 1. Peiris JSM, Lai ST, Poon LLM, et al. Coronavirus as a possible cause of severe acute respiratory syndrome. Lancet 2003;361: 1319-25. SARS and Carlo Urbani Brigg Reilley, M.P.H., Michel Van Herp, M.D., M.P.H., Dan Sermand, Ph.D., and Nicoletta Dentico, M.P.H. On February 28, the Vietnam French Hospital of Hanoi, a private hospital of about 60 beds, con- tacted the Hanoi office of the World Health Or- ganization (WHO). A patient had presented with an unusual influenza-like virus. Hospital officials sus- pected an avian influenzavirus and asked whether someone from the WHO could take a look. Dr. Car- lo Urbani, a specialist in infectious diseases, an- swered that call. In a matter of weeks, he and five other health care professionals would be dead from a previously unknown pathogen. We now know that Hanoi was experiencing an outbreak of severe acute respiratory syndrome (SARS). Dr. Urbani swiftly determined that the small private hospital was facing something unusu- al. For the next several days, he chose to work at the hospital, documenting findings, arranging for sam- ples to be sent for testing, and reinforcing infection control. The hospital established an isolation ward that was kept under guard. Dr. Urbani worked di- rectly with the medical staff of the hospital to strengthen morale and to keep fear in check as SARS revealed itself to be highly contagious and virulent. Of the first 60 patients with SARS, more than half were health care workers. At a certain mo- ment, many of the staff members made the difficult decision to quarantine themselves. To protect their families and community, some health care workers put themselves at great personal risk, deciding to sleep in the hospital and effectively sealing them- selves off from the outside world. In some ways, the SARS outbreak in Hanoi is a story of what can go right, of public health’s com- ing before politics. First-line health care providers quickly alerted the WHO of an atypical pneumonia. Dr. Urbani recognized the severity of the public health threat. Immediately, the WHO requested an emergency meeting on Sunday, March 9, with the Vice Minister of Health of Vietnam. Dr. Urbani’s temperament and intuition and the strong trust he had built with Vietnamese authorities were critical at this juncture. The four-hour discussion led the government to take the extraordinary steps of quar- antining the Vietnam French Hospital, introducing new infection-control procedures in other hospi- tals, and issuing an international appeal for expert assistance. Additional specialists from the WHO and the Centers for Disease Control and Prevention (CDC) arrived on the scene, and Médecins sans Frontières (MSF, or Doctors without Borders) re- sponded with staff members as well as infection- control suits and kits that were previously stocked for outbreaks of Ebola virus. The Vietnam French Hospital has been closed temporarily, and patients with SARS are cared for in two wards of the public Bach Mai Hospital, with the assistance of a team from MSF. No new cases in health care workers have been reported, and the outbreak in Vietnam appears to be contained. By dealing with the outbreak open- ly and decisively, Vietnam risked damage to its im- age and economy. If it had decided to take refuge in secrecy, however, the results might have been cata- strophic. Dr. Urbani would not survive to see the success- es resulting from his early detection of SARS. On March 11, he began to have symptoms during a flight to Bangkok. On his arrival, he told a colleague from the CDC who greeted him at the airport not to approach him. They sat down at a distance from each other, in silence, waiting for an ambulance to assemble protective gear. He fought SARS for the next 18 days in a makeshift isolation room in a Bangkok hospital. Dr. Carlo Urbani died on March 29, 2003. The New England Journal of Medicine Downloaded from nejm.org on May 23, 2014. For personal use only. No other uses without permission. Copyright © 2003 Massachusetts Medical Society. All rights reserved.