Epidemiologic Investigation

Potential close contacts were identified through interviews with the case-patient, his family members, healthcare workers, and other persons. Additional clinical and contact information was obtained through review of medical records. “Close contact” exposures included any amount of time spent within 3 feet of the patient or 30 minutes or longer within 3 to 10 feet. Since evidence suggests that SARS-CoV is primarily transmitted by means of large respiratory droplets, usually spread within a 3-foot radius, we focused on contact within this range (16). Thirty minutes within the patient’s immediate care area (3–10 feet) was chosen arbitrarily to divide shorter and longer exposures.

Contacts included persons exposed to the patient before and after his diagnosis as a probable SARS patient. Contacts were grouped according to sites of principal exposure: the term “healthcare workers” refers to employees or contractors of a healthcare facility, “healthcare -related contacts” includes non–healthcare worker contacts exposed in a healthcare setting, “household contacts” includes immediate family members, whether they resided in the same household or not, and “community contacts” includes persons exposed in other settings. Public health personnel, using standard data collection instruments, interviewed contacts regarding their type and duration of contact with the patient, use of personal protective equipment, and clinical symptoms after contact. Direct, unprotected contact with the patient’s skin (i.e., without gloves) was defined as skin-to-skin contact, and unprotected contact with inanimate objects likely to have been touched by the patient, such as bedrails and clothing, was defined as skin-to-object contact.

Contacts were defined as prediagnosis or postdiagnosis contacts. Prediagnosis contacts were those exposed to the case-patient after his onset of symptoms (April 3) but before the patient’s diagnosis of probable SARS (April 14). Postdiagnosis contacts were those exposed only after the diagnosis was made and infection control precautions were in effect. A convenience sample of postdiagnosis contacts was selected because strict infection control procedures had already been instituted, with all contacts wearing personal protective equipment; thus, unprotected exposures were not anticipated. Of the 32 persons with postdiagnosis exposure exclusively, 15 healthcare workers were selected for epidemiologic and laboratory evaluation.

Biologic Specimen Collection

Serum, whole blood (collected into a tube containing EDTA), oropharyngeal swab (swab of posterior pharynx), stool, and urine samples were requested from the case-patient twice weekly until day 21 after symptom onset and weekly for 2 additional weeks. In addition, a single nasopharyngeal swab specimen, nasal aspirate, and sputum sample were collected from the case-patient while he was hospitalized. The first set of specimens requested from his prediagnosis contacts included serum, whole blood, nasopharyngeal and oropharyngeal swab specimens, stool, and urine. Thereafter, specimens (serum, whole blood, oropharyngeal swab, and stool) were requested from prediagnosis contacts weekly until at least 22 days after the most recent exposure to the case-patient. Healthcare workers with postdiagnosis exposure submitted a single set of convalescent-phase specimens (>21 days after the last exposure), including serum, whole blood, and an oropharyngeal swab. Nasopharyngeal and oropharyngeal swab specimens were collected by using Dacron swabs with nonwooden handles. Swabs were immediately placed into viral transport medium and placed on ice. All specimens were stored at 4°C and shipped within 72 hours of collection to the Centers for Disease Control and Prevention (CDC).

Two postdiagnosis healthcare workers, in whom fever developed after they were exposed to the case-patient, provided weekly specimens rather than a single set. One prediagnosis healthcare-related contact participated until 22 days after exposure but did not provide serum or whole blood specimens, and four prediagnosis contacts (2 healthcare workers and 2 healthcare-related contacts) declined further participation after specimen collection at 8, 11, 11, and 21 days after exposure, respectively.

Environmental Specimen Collection

Sterile Dacron swabs with nonwooden handles were moistened with sterile saline or viral transport medium and rolled over environmental surfaces, including toilet and sink surfaces and other commonly touched items (e.g., door handles, telephones, remote controls, and toiletries) and placed in viral transport medium. Twenty environmental swab samples were collected from the patient’s hospital room during his hospitalization (day 17 after illness onset), and 12 were collected from his home bedroom and private bathroom 3 days after hospital discharge (day 21 after illness onset). These were stored and shipped to CDC at 4°C.

Laboratory Testing

To test for evidence of infection with SARS-CoV, total anti–SARS-CoV serum antibody was measured by enzyme-linked immunosorbent assay (ELISA) and indirect fluorescent antibody test (17). Reverse transcription–polymerase chain reaction (RT-PCR) was performed on nasopharyngeal and oropharyngeal swabs and stool and urine specimens; results were confirmed in separate CDC laboratories, with both negative and positive controls (17,18). Quantitative RT-PCR on stool specimens was conducted by using the TaqMan assay and standard curves generated from synthetic RNA transcripts (17). Viral culture in Vero E6 cells was performed on all RT-PCR–positive specimens (17).

Human Participants

This investigation was conducted as part of CDC’s public health response to the SARS outbreak. Informed consent was obtained from the case-patient and contacts before epidemiologic information was obtained and biologic specimens were collected.

Statistical Analysis

Due to the non-Gaussian distribution of the data, the Wilcoxon rank-sum test was used to compare median durations of contact between different groups of persons. Prevalences of different types of exposures between the groups were compared by using Fisher exact test.

Clinical History and Laboratory Findings for the Case-Patient

Figure 1

Figure 1. Timeline: severe acute respiratory syndrome (SARS) case-patient symptoms and total daily number of contacts from date of symptom onset to date of hospital discharge. Contacts indicated regardless of their subsequent participation...

After traveling by automobile to an event held in Toronto on March 29 and 30, the previously healthy patient had onset of myalgias, subjective fever, chills, and diaphoresis on April 3 (Figure 1). Diarrhea developed on April 5, and the patient sought medical care at the emergency department of hospital A on April 6. The patient had a temperature of 38.2°C (100.7°F) and was discharged with a diagnosis of acute viral syndrome; no diagnostic testing was performed. During this emergency department visit, the patient did not report recent travel to Toronto to healthcare providers. By April 10, despite taking oral amoxicillin for 3 days (initiated after telephone consultation with his primary care physician), a dry cough developed, which prompted him to visit his primary care physician. His physician referred him to an outpatient laboratory for phlebotomy and to hospital B for chest radiography; findings on the radiograph were normal, and the patient was sent home.

On April 14, the patient went to the emergency department of hospital B with dehydration, worsening cough, and severe shortness of breath. Within 2.5 hours of arrival, a diagnosis of SARS was suspected on the basis of a full travel history and new radiographic evidence of pneumonia. The patient was admitted to an airborne-infection (negative-pressure) isolation room, and the hospital instituted contact and airborne precautions for all healthcare workers in contact with the patient, restricted visitation to this patient, and immediately notified public health authorities. Serum samples collected on April 14 (day 11 of illness) demonstrated antibodies to SARS-CoV. Admission vital signs included a temperature of 37.7°C (99.9°F) and oxygen saturations of 90%–91% on room air. The patient was given supportive care (including 2 days of supplemental oxygen, inhaled fluticasone propionate/salmeterol twice daily, and antimicrobial drugs (levofloxacin for pneumonia and metronidazole for diarrhea associated with laboratory-confirmed Clostridium difficile infection). His highest documented temperature while hospitalized was 38.1°C (100.6°F) on April 15. After the patient was hospitalized for 4 days, his fever and systemic symptoms resolved, and he was discharged on April 21 (hospital day 7) with a persistent but improving cough. He did not require aerosolized nebulizer treatments, intubation, or admission to an intensive care unit during his hospitalization.

Figure 2

Figure 2. Clinical specimens collected and laboratory results for Pennsylvania severe acute respiratory syndrome (SARS) case-patient, April 2003. Symbols of specimens and method of testing: serum anti–SARS-CoV antibody, circles; stool RT-PCR; squares; urine...

The case-patient’s serum specimens from days 11 to 32 after illness onset demonstrated anti–SARS-CoV antibodies (Figure 2). Additional analysis showed an increase in antibody titer over time (19). All respiratory specimens and the only urine sample tested negative by RT-PCR for SARS-CoV. However, serial stool specimens collected on days 14, 18, 21, and 26 after the onset of illness were positive by RT-PCR. Quantitative PCR showed the copy number in the first collected stool to be 16- to 40-fold higher than that in all subsequent stools (19). Viral cultures of all stools and respiratory specimens were negative for SARS-CoV, and all environmental specimens were negative by RT-PCR for SARS-CoV.

Epidemiologic and Laboratory Results for Contacts

The principal potential exposure sites that were investigated included sites for healthcare worker and healthcare-related contact exposures (emergency department of hospital A; primary care physician’s office; referral phlebotomy laboratory; and emergency department, radiology suite, and inpatient facility of hospital B), the patient’s home, and community settings in which the patient reported having had close contacts.

Prediagnosis Contacts

Thirty-four potential prediagnosis contacts were identified, and questionnaires were collected from 26 (76%) of them. The eight remaining potential prediagnosis contacts, who did not complete questionnaires, included seven healthcare-related contacts (six who were present in a laboratory waiting room at the same time as the case-patient and one radiology staff member) and one community contact (a retail salesperson). Of these eight persons, two could not be contacted, five did not complete more detailed interviews but did not recall specific interaction with the patient or report any subsequent illness, and one reported brief contact with the patient with no subsequent symptoms and declined to answer further questions.

The 26 prediagnosis contacts who completed questionnaires included 4 household contacts (15%), 17 healthcare workers (65%), and 5 others (19%), including 4 healthcare-related contacts (4 persons in a waiting room or curtained area in the emergency department) and 1 community contact (a bank teller) (Table). The median age of prediagnosis contacts was 41.3 years (range 15.7–90.1); the only 2 contacts over age 65 were healthcare-related contacts.

Figure 3

Figure 3. Duration of exposure for close contacts within 3 feet on the three dates when the case-patient with severe acute respiratory syndrome sought medical care. Four contacts (three household contacts and one...

Of these 26 persons, nearly all (92%) had contact with the patient during the 3 days when he sought medical care (Figures 1, 3). All household contacts and healthcare workers with prediagnosis contact had close unprotected exposures (within 3 feet), compared with 40% of the other contacts; this finding was significantly different only for healthcare workers (p = 0.006; p = 0.17 for household contacts) (Table). However, household contacts had the longest median duration of exposure per person, 60 times longer than the median duration per person among prediagnosis healthcare workers (459 vs. 7.5 minutes, p = 0.04) and 15 times longer than among other contacts (459 vs. 30 minutes, p = 0.008). Household contacts and healthcare workers had similar degrees of skin-to-skin contact (50% vs. 53%, p = 1.00) and skin-to-object contact (100% vs. 71%, p = 0.53). The patient and household contacts attempted to limit interactions throughout his illness and began wearing surgical masks when they interacted after April 9.

All contacts were monitored for fever and respiratory symptoms during the 10 days after exposure to the case-patient. Eleven (42%) of the 26 prediagnosis contacts reported fever and/or lower respiratory tract symptoms (defined as cough, wheezing, or shortness of breath/difficulty breathing) during the surveillance period. Of the 26, 1 (4%) reported fever alone, 9 (35%) reported respiratory symptoms alone, and 1 reported both. The person with both fever and respiratory tract symptoms was a household contact who reported sore throat and cough before contact; fever developed after contact, thus meeting the CDC clinical case definition for a suspected SARS case (9,20). Seven (41%) of 17 healthcare workers with prediagnosis contact were furloughed from work for 3 to 10 days due to unprotected close contact or the presence of respiratory symptoms. Four (57%) of these persons had lower respiratory tract symptoms, and three (43%) were asymptomatic or had only mild symptoms (sore throat, headache, or rhinorrhea).

Prediagnosis contacts provided a total of 86 serum and whole blood samples, 90 oropharyngeal swabs, 25 nasopharyngeal swabs, 18 stool samples, and 4 urine specimens (Table). The household contact who met the suspected SARS case definition provided a single nasopharyngeal swab, stool, and urine samples, and acute- and convalescent-phase (37 days after contact) serum specimens, whole blood samples, and oropharyngeal swabs. The other contact with fever provided a single nasopharyngeal swab and stool sample and three oropharyngeal swabs, serum specimens, and whole blood samples (up to 22 days after contact). The median time after contact to collection of the last serum specimen was 28 days (range 8–37). All specimens tested negative for SARS-CoV.

Postdiagnosis Contacts

Some contacts had unprotected exposures within 3 feet on the day SARS was diagnosed in the case-patient; the most prolonged of these were 210 minutes for a household contact and 30 minutes, including skin-to-skin contact, for a community contact (Figure 3). However, nearly all contacts were protected after diagnosis (Figures 1, 3; Table). The sample of 15 postdiagnosis healthcare workers was protected with fit-tested N95 respirators, gowns, and gloves (goggles were added on day 2 of hospitalization). Postdiagnosis healthcare workers had a median age of 39.1 years (range 24.6–51.7). Despite much longer median durations of exposure compared with those of the prediagnosis healthcare workers (110 vs. 7.5 minutes/person, p<0.005; Table), postdiagnosis healthcare workers had only two unprotected close contacts, one failure to wear a gown, and one failure to wear an N95 respirator and gloves during skin-to-skin contact.

After contact with the patient, two (14%) postdiagnosis healthcare workers reported fever. One of these persons also reported a cough 2 days after exposure to the case-patient and, therefore, met the clinical case definition for suspected SARS (9,20). This person was admitted to the hospital for 1 night with a diagnosis of respiratory syncytial virus infection (antigen-positive nasal aspirate) and asthma exacerbation. Neither of these symptomatic postdiagnosis healthcare workers had breaches in personal protection equipment. All specimens from postdiagnosis healthcare workers tested negative for SARS-CoV, including specimens from both contacts with fever, each of whom provided a single nasopharyngeal swab and weekly oropharyngeal swabs, serum specimens, and whole blood samples (up to 27 and 28 days after contact).