Severe Acute Respiratory Syndrome

1.
2.
3.
4.
5.

Michael D. Christian1,
Susan M. Poutanen2,3,
Mona R. Loutfy5,
Matthew P. Muller2,4, and
Donald E. Low2,3,4
+Author Affiliations
1. 1Toronto General Hospital, University Health Network, Toronto, Ontario
2. 2Toronto Medical Laboratories & Mount Sinai Hospital Department of
Microbiology, Toronto, Ontario
3. 3Departments of Laboratory Medicine and Pathobiology, Toronto, Ontario
4. 4Medicine, University of Toronto, Toronto, Ontario
5. 5Department of Medicine, McGill University, Montreal, Quebec
1.
Reprints or correspondence: Dr. Donald E. Low, Mount Sinai Hospital, 600
University Ave., Rm. 1487, Toronto, Ontario, Canada M5G 1X5 ([email protected]).
Next Section
Abstract
The first cases of severe acute respiratory syndrome (SARS) occurred in China in
November 2002. The agent causing this illness has been identified as a novel coronavirus,
SARS-coronavirus. Since its introduction <1 year ago, this virus has infected 8098 people
in 26 countries, killing 774 of them. We present an overview of the epidemiology, clinical
presentation, diagnosis, and treatment of SARS based on the current state of knowledge
derived from published studies and our own personal experience.
On 11 February 2003, the Program for Monitoring Emerging Diseases
(http://www.promedmail.org) reported that, since November 2002, an unidentified agent
had caused some 300 cases of pneumonia in persons in the south of China. On 12 March
2003, the World Health Organization (WHO) issued a global alert regarding these and
similar cases in Hong Kong and Vietnam. This clinical syndrome subsequently became
known as severe acute respiratory syndrome (SARS). Since then, 8098 people in 26
countries have had probable SARS diagnosed, 774 of whom have died (figure 1), yielding
a global case-fatality rate of 10% [1, 2]. On 5 July 2003, the WHO reported that the last
known human chain of transmission of SARS had been broken [3].

View larger version:

In this page

Figure 1

In a new window
Download as PowerPoint Slide

Probable cases of severe acute respiratory syndrome reported worldwide between 1

November 2002 to 31 July 2003 [1]. aIncludes France, Germany, Ireland, Italy, Romania,
Spain, Sweden, Switzerland, and United Kingdom. bIncludes Macao. , Date of onset of
first probable case; , Date of onset of last probable case.
A newly discovered coronavirus (SARS-CoV) has been identified as the cause of SARS
[47]. SARS-CoVlike viruses have been detected in Himalayan palm civets and a
raccoon-dog in a market in southern China, suggesting that the origin of SARS-CoV may
have been from these or other wild animals [8]. Given the possibility that human or animal
reservoirs of SARS-CoV may still exist and that SARS may have a seasonal predilection,
there is concern that SARS may return in upcoming respiratory seasons. WHO guidelines
emphasize the need for all countries to remain vigilant and to maintain their capacity to
detect and respond to the potential reemergence of SARS [9].
Previous SectionNext Section
Epidemiology
SARS remained isolated in China from November 2002 until 21 February 2003, when a
physician with SARS traveled from Guangdong province to a hotel in Hong Kong, infecting
10 other guests [9]. The movements of these 11 individuals resulted in the spread of SARS
worldwide and sparked all of the major epicenters outside of China [1] (figure 2).

View larger version:

In this page

In a new window
Download as PowerPoint Slide

Figure 2
Transmission of severe acute respiratory syndrome by a single individual from Guangdong
province, China, to Hotel M in Kowloon, Hong Kong, and subsequently to countries around
the world via air travel by infected hotel guests [9].
The rate of spread of an epidemic and whether it is self-sustaining depend on the basic
reproduction number (R0). R0 is defined as the average number of secondary cases
generated by 1 primary case in a susceptible population [10]. This quantity determines the
potential for an infectious agent to start an outbreak, the extent of transmission in the
absence of control measures, and the ability of control measures to reduce spread. During
the course of an epidemic, Rt the effective reproduction number, decreases in comparison
with R0 as a result of the depletion of susceptible persons in the population, death or
recovery with subsequent immunity, and the implementation of specific control measures.
To stop an outbreak,Rt must be maintained below 1. Mathematical modeling of the early
phase of the Singapore and Hong Kong outbreaks, before the institution of control
measures and during which time it was occurring primarily in the hospital setting,

estimated that the R0 was 2.23.7, indicating that the virus is moderately infective [11, 12].
The attack rate for SARS-CoV ranges from 10.3% to 60% or 2.4 to 31.3 cases/1000
exposure-hours, depending on the clinical setting and the unit of measurement [13]. A
significant limitation of these calculations is that these data are based on diagnoses made
with a clinical case definition. Reanalysis will be required once the results of
seroprevalence studies are completed and will provide a more accurate estimate of R0.
Previous SectionNext Section
Mechanisms and Routes of Transmission
SARS-CoV has been isolated in sputum samples, nasal secretions, serum specimens,
feces samples, and bronchial washings [5, 14]. Evidence suggests that SARS-CoV is
transmitted by contact and/or droplets [6, 15] and that the use of any mask (surgical or
N95) significantly decreases the risk of infection [16]. However, there are cases that defy
explanation based on these modes of transmission, suggesting that alternative modes of
transmission may also exist [13, 17]. SARS-CoV remains viable in feces for days [18], and
the outbreak at the Amoy Gardens apartments highlights the possibility of a fecal-oral or
fecal-droplet mode of transmission [19,20]. A number of cases occurred in health care
workers wearing protective equipment following exposure to high-risk aerosol- and dropletgenerating procedures, such as airway manipulation, administration of aerosolized
medications, noninvasive positive pressure ventilation, and bronchoscopy or intubation
[17, 21, 22]. When intubation is necessary, measures should be taken to reduce
unnecessary exposure to health care workers, including reducing the number of health
care workers present and adequately sedating or paralyzing the patient to reduce cough.
Updated infection control precautions for patients who have SARS are available from the
Centers for Disease Control and Prevention (CDC) at
http://www.cdc.gov/ncidod/sars/index.htm.
Currently, epidemiological evidence suggests that transmission does not occur before the
onset of symptoms or after symptom resolution, even though shedding of SARS-CoV in
stool has been documented by RT-PCR for up to 64 days after the resolution of symptoms
[23]. A small group of patients appear to be highly infectious and have been referred to as
super-spreaders [24]. Such events appear to have played an important role early in the
epidemic. Possible explanations for their enhanced infectivity include the lack of early
implementation of infection-control precautions, higher SARS-CoV load, and larger
amounts of respiratory secretions.
Previous SectionNext Section
Clinical Disease
Case definition. Case definitions of SARS based on clinical and epidemiologic data were
developed during the outbreak. Although these definitions were epidemiologically useful,
Rainer et al. [25] have shown that they had a low sensitivity for diagnosis in patients early
in disease (sensitivity, 26%; specificity, 96%), underscoring the importance of a rapid,
accurate diagnostic test. Since then, the CDC has developed updated SARS surveillance
case definitions based on clinical, epidemiologic, and laboratory criteria [26]. The WHO
has similar updated definitions [27].
Presentation. The typical incubation period of SARS ranges from 2 to 10 days but may
occasionally be as long as 16 days [21, 28]. The frequencies of symptoms at the onset of
disease are summarized in table 1. The prodrome includes influenza-like symptoms, such
as fever, myalgias, headache, and diarrhea [21, 28]. Fever can vary from low to high grade
and can occasionally be absent at presentation, particularly in older patients. The typical
respiratory phase starts 27 days after the prodrome and can be associated with watery

diarrhea [14, 21, 28]. The early respiratory stage includes a dry, nonproductive cough and
mild dyspnea. Early-phase chest radiographs often show subtle peripheral pulmonary
infiltrates that can be more readily detected as consolidations having a ground-glass
appearance with high-resolution CT of the lung [29, 30]. Atypical presentations of the
disease have been described elsewhere [31,32], including cases involving fever but no
respiratory component [33]. Asymptomatic cases have also been described, but only in
small numbers [34]. Of interest, the disease has been rare in children and, when present,
has appeared to be milder [33, 35].

View larger version:

In this page

In a new window
Download as PowerPoint Slide

Table 1
Summary of clinical findings of severe acute respiratory syndrome at admission to the
hospital.
Spectrum of disease. After the onset of disease, cases may progress to a mild variant of
the disease characterized by mild respiratory symptoms with fever or a cough variant
characterized by persistent intractable cough. However, most commonly, cases progress
to a moderate-severe variant characterized by a more serious later respiratory phase with
dyspnea on exertion or at rest and hypoxia. This later respiratory phase typically occurs 8
12 days after the onset of symptoms (table 2) [14, 21,28]. In 10%20% of hospitalized
patients, persistent or progressive hypoxia results in the requirement of intubation and
mechanical ventilation [28,36, 37]. Among patients developing respiratory failure,
intubation is required at a median of 8 days after onset of symptoms [35, 36]. Subtle but
progressive decreases in oxygen saturation are often indicative of impeding respiratory
failure and should trigger more-intensive monitoring and preparation for intubation under
controlled circumstances. Typically, the respiratory phase lasts 1 week. The recovery
phase begins 1418 days after the onset of symptoms.

View larger version:

In this page

In a new window
Download as PowerPoint Slide

Table 2
Duration of clinical phases of the mild and moderately severe variants of severe acute
respiratory syndrome.
Clinical outcome. The case-fatality rate during recent outbreaks was 9.6% (range, 0%
40%) [1]. Advanced age is the most important risk factor for death: patients aged >60
years have a case-fatality rate of 45% [14, 28]. Other risk factors for death include
diabetes mellitus and hepatitis B virus infection [14, 21, 28, 36, 37]. Little data exist
regarding the long-term morbidity of SARS, although preliminary studies suggest that the
psychological impact of the disease is considerable [38, 39].
Previous SectionNext Section
Laboratory Diagnosis
Sensitive and specific tests for detection of SARS-CoV that can yield results within hours
of patient presentation are urgently needed. Many tests have been developed, including
some that are now marketed commercially, but none have yet achieved this goal. The
early diagnosis of SARS is based on recognition of epidemiological linkages; the presence
of typical clinical, laboratory, and radiographic features; and the exclusion of other
respiratory pathogens. None of these features of SARS are specific, however, and
diagnosis should be confirmed by SARS-CoVspecific microbiological and serological
studies, although initial management will continue to be based on a clinical and
epidemiological assessment of the likelihood of SARS-CoV infection.
Routine laboratory tests. At the time of presentation, routine hematological and
biochemical test findings are frequently abnormal. In particular, lymphopenia,
thrombocytopenia, an absence of neutrophilia, elevated lactic dehydrogenase levels, and
elevated creatine phosphokinase levels have been observed [14, 21, 28]. However, on the
basis of preliminary analysis, when compared against the laboratory features of other
causes of community-acquired pneumonia, none of these tests alone are sufficient to
significantly alter the pretest probability of SARS [40].
Testing for SARS-CoV. Serologic assays for SARS-CoV include immunofluorescent
assays, ELISAs, and Western blot assays. On the basis of the limited studies completed to
date, IgM detection is delayed by 1 week after the onset of symptoms, and the mean
time to IgG seroconversion is estimated to be between 20 and 26 days. As a result,
serologic testing at presentation is not useful as a strategy for rapid diagnosis [14, 41, 42].
Assays for culturing SARS-CoV in cell lines and for its rapid detection by RT-PCR from
clinical specimens are available. The sensitivity of culture is lower than that of RT-PCR
[43]. The specificity of the RT-PCR assays can be assured by using specific probes,
sequencing the RT-PCR product, or completing a second RT-PCR assay using primers
amplifying a different genome region. Quality-control measures should be in place to
prevent false-positive results due to laboratory contamination. The analytic sensitivity of
these assays has been shown to be high, with reproducible detection limits of 10 copies of
viral RNA [44]. However, on the basis of the results of first-generation assays, the clinical
sensitivity of SARS-CoV RT-PCR has been estimated to be as low as 50%, depending on
the type of specimens tested and the timing of collection relative to the onset of symptoms
of SARS [6, 42]. Specimens that have the highest proportion positive for SARS-CoV using
first-generation assays include nasopharyngeal swabs, nasopharyngeal aspirates, throat
swabs, and stool (highest yield) [14, 42, 43]. Specimen obtainment 10 days from
symptom onset is associated with the highest yield for all specimen types, which correlates
with the timing of peak virus loads [14]. Optimization of SARS-CoV RT-PCR assays with
regard to the targeted genomic region, the timing of specimen obtainment, the type of

specimens, and the extraction methodology is ongoing. A recent report describes a second
generation assay with a clinical sensitivity of 80% using a modified extraction method on
nasopharyngeal aspirates obtained in the first 3 days of illness [45].
A summary of samples and tests to investigate possible SARS appears intable 3.
Coinfections with SARS-CoV and other infectious agents can exist, and, as a result,
finding an infectious agent other that SARS-CoV should not be used to rule out SARS [46].

View larger version:

In this page

In a new window
Download as PowerPoint Slide

Table 3
Suggested tests for patients presenting with a high pretest probability of severe acute
respiratory syndrome (SARS).
Previous SectionNext Section
Treatment
At present, there is not sufficient evidence to recommend any specific therapy for the
treatment of SARS. Because SARS cannot be easily distinguished from other causes of
pneumonia, patients who are suspected of having SARS who have pulmonary infiltrates
should receive appropriate antibiotic coverage [47]. Respiratory failure is the primary
cause of acute morbidity and mortality due to SARS-CoV infection and occurs in 20%
25% of cases [21, 28, 37]. When mechanical ventilation is required, a lung protective
ventilation strategy should be used based on an analogy to data for the treatment of acute
respiratory distress syndrome (ARDS) and the current intensive care unit experience
managing SARS [36, 37, 48]. In fact, barotrauma appears to be one of the most frequent
complications of severe SARS-CoV infection, with pneumothorax and/or
pneumomediastinum occurring in 20%34% of ventilated patients, a rate that is much
higher than the rate of 2.5% observed in a large study of ARDS [36, 37, 48].
Antiviral therapy. Antiviral agents used in the therapy of SARS include ribavirin, IFN-,
and lopinavir-ritonavir. Ribavirin is a nucleoside analogue with in vitro activity against a
number of RNA and DNA viruses, including some animal coronaviruses [49]. Ribavirin was
widely used for the treatment of SARS. Initial reports noted improvement in surrogate
markers of outcome, such as resolution of fever and improvement in oxygenation and
radiographic appearance [15, 21, 50]. These studies were not controlled, and most
patients also received corticosteroids [15, 21, 50]. Other reports failed to identify
improvement with ribavirin [28, 51], and one report identified a high frequency of adverse
events among patients treated with high-dose ribavirin, including severe hemolysis (in 49%
of patients) [28]. In vitro testing of SARS-CoV indicated that ribavirin does not have activity
against this virus at clinically achievable concentrations [52]. Postmortem findings for
some patients demonstrated that high virus loads persisted despite treatment with ribavirin
[53].

IFNs are cytokines with well-described antiviral activity [54]. IFNs, particularly IFN-, inhibit
SARS-CoV in vitro [55]. An open-labelled study using IFN-alfacon-1 and high-dose
methylprednisolone demonstrated more-rapid improvement in radiographic appearance
and oxygenation in recipients, compared with a historic cohort that received a lower dose
of corticosteroids alone [56]. A complex 4-arm trial examining ribavirin and IFN and
differing doses of corticosteroids also demonstrated improvement in surrogate end points,
such as radiographic appearance, but these improvements only occurred in the IFN
recipients who also received high-dose corticosteroids [57].
Lopinavir-ritonavir is a combination drug consisting of 2 protease inhibitors with proven
efficacy in the treatment of HIV. Lopinavir-ritonavir was studied in a nonrandomized open
label study in Hong Kong as initial and rescue therapy for SARS. It was added to local
standard therapy consisting of ribavirin and corticosteroids, and, when used as initial
therapy, recipients had a significant reduction in the overall death rate and intubation rate,
compared with a matched control group who received standard treatment alone. The
control group, however, had lower rates of steroid use at lower mean doses, making
definitive conclusions difficult [58].
Anti-inflammatory therapy. Anti-inflammatory or immunomodulatory therapies include
corticosteroids, intravenous immunoglobulin (IVIG), and convalescent-phase serum and
plasma exchange. Corticosteroids were widely used for SARS therapy. Preliminary results
demonstrate decreasing virus loads and increasing antibody titers during the second week
of illness, at a time when the respiratory disease typically progresses [14]. These results
suggest that lung damage in patients with SARS-CoV infection may be immune mediated
and provides the rationale for corticosteroid therapy. Pathological findings are consistent
with cytokine dysregulation and provide further support for the theory that lung damage is
immune mediated [52]. Initial case reports described resolution of fever and improvements
in oxygenation and radiographic appearance in some patients treated with ribavirin and
corticosteroids [59]. Subsequently, clinicians noted that cases in many patients progress
despite receiving treatment with corticosteroids, and higher doses or pulsed steroid
regimens were required as rescue therapy [60, 61]. A trial comparing early use of pulsed
versus nonpulsed corticosteroids did not note any difference in the requirement for
ventilation or mortality, but it did reveal improvements in oxygenation and radiographic
appearance [62].
Concerns with the use of prolonged, high-dose, and pulsed corticosteroid regimens in the
treatment of a new viral infection have been raised [62]. Pathological studies have
detected high virus loads in patients who died >50 days into their illness and suggest that
persistent viral replication is occurring and likely contributes to the pathophysiology of lung
damage in SARS-CoV infection [53]. The use of corticosteroids could potentially increase
or prolong viral replication and thereby worsen disease. Corticosteroids are also
associated with a number of well-known adverse outcomes including immunosuppression
and increased susceptibility to opportunistic pathogens, particularly those causing invasive
fungal infections [63, 64]. In one multivariate analysis of 218 patients with SARS, use of
pulsed corticosteroid therapy was strongly associated with mortality, although the results
are difficult to interpret, because the sickest patients typically received pulsed
corticosteroids as salvage therapy [65]. There have also been reports from China of
avascular necrosisa well-recognized complication of steroid therapyoccurring in
patients who recovered from SARS [66]. IVIG consists of pooled antibodies from multiple
donors. Although IVIG would not be expected to contain antibodies directed against
SARS-CoV because of the low rate of background immunity in the population, it has been
demonstrated to have other immunomodulatory properties and may down-regulate
cytokine expression [67]. Convalescent-phase plasma samples have been used to treat
SARS in Hong Kong and China and may be of value because, unlike standard IVIG

preparations, high levels of antiSARS-CoV antibodies would be present [68]. Finally,

plasma exchange was used as salvage therapy in Hong Kong, but no data exist with which
to assess its efficacy [69].
Previous SectionNext Section
Conclusions
SARS is a deadly new infectious disease with the ability to spread from person to person
and from country to country via international air travel. Despite the subsequent rapid
spread of this virus worldwide, traditional public health measures were able to contain and
control this outbreak. Because SARS-CoV causes a nonspecific clinical illness, diagnosing
and controlling this disease in the future will require the development of rapid accurate
tests. There is an urgent need to develop means of performing clinical trials that evaluate
treatment regimens for SARS, as well as other new diseases, especially in outbreak
situations.
Previous SectionNext Section
Acknowledgments
We would like thank Yuan Zhang and Raymond Chow, for producing the graphics in this
article, and Alice Au Yeung, for her assistance with the manuscript preparation.