Participants

We targeted 103 patients who met the Centers for Disease Control and Prevention’s (CDC’s) surveillance case definition for probable SARS (14). Of these patients, 7 (7%) with laboratory-confirmed SARS-CoV infection (antibodies to SARS-CoV were detected) were enrolled; 19 (18%), including 1 confirmed SARS case-patient, declined participation; and 77 (75%) were excluded for various reasons (negative for SARS-CoV antibody at ≥21 days after illness onset, a confirmed alternative diagnosis, or foreign citizen not residing in the United States). The household contacts of seven laboratory-confirmed case-patients were also enrolled. Household contacts were defined as persons who had lived in the same household with SARS case-patients during their illness. All participants provided informed consent.

Timeline for Follow-up Visits

Follow-up visits were scheduled twice a week for the first 3 weeks after illness onset and then once a week for 2 weeks thereafter. If a case-patient was first enrolled after week 5 of illness, then a follow-up visit was made as soon as feasible after enrollment. Some case-patients were enrolled at >5 weeks after illness onset and, therefore, were followed up for >10 weeks. For household contacts, visits were scheduled once weekly for a period of 4 weeks after initial exposure to the case-patient. A single follow-up visit was scheduled if the household contact was enrolled >4 weeks after initial exposure to the case-patient.

Clinical and Epidemiologic Data

At the initial visit with the SARS case-patients, we collected data on demographics, date of illness onset, clinical symptoms, and exposure history. At the initial visit with household contacts, we gathered data on any illness they had had since their exposure to the case-patient and on the types and patterns of exposure (e.g., sleeping in the same room at night, daily contact within <3 feet, and direct skin-to-skin contact, such as kissing or hugging, with case-patients). At each subsequent visit, we collected information on any symptoms experienced by case-patients or household contacts since their previous visit, including symptoms during the current visit.

Clinical Specimens

Specimens collected as a part of the diagnostic work-up were available for this investigation, and at each postenrollment visit, participants were asked to provide whole-blood, serum, stool, urine, nasopharyngeal, and oropharyngeal swab specimens. We obtained 1–10 mL of blood from adults and 0.5–5 mL of blood from children <3 years old by venipuncture or finger stick. Clotted blood was centrifuged, and serum was separated before being shipped to CDC for testing. Similar volumes of whole blood were collected in a tube containing EDTA. Nasopharyngeal and oropharyngeal samples were collected by use of a single Dacron swab with a nonwooden shaft; the swab was then placed in a sterile vial containing 2 mL of viral transport medium. Stool specimens were collected in a sterile container and sealed. Participants provided a 50-mL clean-catch collection of urine in a sterile urine cup. Specimens were processed and stored according to CDC laboratory biosafety guidelines (15). All specimens were stored at 4°C for a maximum of 72 h and shipped on ice to the CDC laboratory. If shipping within 72 h was not feasible, specimens were stored at –70°C and then shipped.

Laboratory Methods

To detect SARS-CoV in stool, urine, and respiratory specimens, we performed reverse transcriptase–polymerase chain reaction (RT-PCR), using primers targeted to the polymerase and nucleocapsid genes of the SARS-CoV genome, as described elsewhere (2, Emery et al., unpub data). Stool samples were prepared as 10% extracts in Tris-HCl buffer before isolation of total nucleic acid for RT-PCR testing. To quantify the virus load in respiratory and stool specimens, quantitative RT-PCR was performed using the TaqMan assay and standard curves generated from synthetic RNA transcripts (S. S. Monroe & R. S. Beard, unpub. data). Previously described culture techniques (2) were used to isolate SARS-CoV from specimens. To determine the S and N gene sequences of SARS-CoV, a set of 10 overlapping RT-PCR products, which cover the entire open reading frames of the S (8 products) and N (2 products) genes, were generated by using the SuperScript One-Step RT-PCR with Platinum Taq (Invitrogen, Carlsbad, CA) and sequenced by using 16 (S gene) or 7 (N gene) sequencing primers (S. Tong et al, unpub. data). Serum specimens were tested for SARS-CoV–associated antibodies by use of an enzyme-linked immunosorbent assay and an indirect fluorescent antibody test, using previously described methods (2). Serum specimens were considered positive only if results for both tests were positive using predetermined cut-offs (2).

Follow-up Findings

Five of seven enrolled case-patients provided data on residual symptoms. Three case-patients reported shortness of breath that persisted at least until days 50, 56, and 62, respectively, after onset of fever. Two case-patients reported residual coughing: case-patient 4 reported a dry cough until day 50 and case-patient 2 reported a productive cough until day 56 after onset of fever. These symptoms had been reported during the acute phase of each case-patient’s illness. Wheezing developed in one case-patient without a previous history of respiratory disease at day 11 of illness and persisted at least until day 46. No data were available to characterize the progression of symptoms over time.

Figure

Figure. Detecting severe acute respiratory syndrome–associated coronavirus (SARS-CoV) RNA by reverse transcriptase–polymerase chain reaction (RT-PCR) and SARS-CoV antibodies by enzyme-linked immunosorbent assay (ELISA) in clinical specimens from seven confirmed SARS case-patients, United...

Of 41 respiratory specimens obtained from seven case-patients (Table 1), 4 (10%) were sputum samples from two case-patients (1 from case-patient 5 and 3 from case-patient 7). SARS-CoV was detected by both RT-PCR and viral culture in the sputum sample of case-patient 5, which was collected at day 14 after illness onset (Figure). All other respiratory specimens, including seven nasopharyngeal and oropharyngeal swab samples collected during the first 2 weeks of illness from five case-patients, tested negative by RT-PCR.

A total of 14 stool specimens were obtained from seven case-patients: two patients provided 4 samples each, one patient had 2 samples, and four had 1 sample each. SARS-CoV RNA was detected in five specimens, all of which came from two case-patients (one specimen from case-patient 6 and four specimens from case-patient 7) (Figure). The single positive stool specimen from case-patient 6 was obtained 19 days after onset; his subsequent stool specimens (collected at days 23, 32, and 44) tested negative for SARS-CoV by RT-PCR. The first stool specimen from case-patient 7 was collected on day 14 of illness; viral RNA was detected in all four of his stools, including the last one, which was collected at day 26. SARS-CoV was not isolated by culture from any of the RT-PCR–positive stool specimens.

The highest concentrations of SARS-CoV were detected in sputum from case-patient 5 (43 million copies per gram of specimen) and in the day-14 stool from case-patient 7 (37 million copies per gram of specimen) (Table 2). After day 14 of illness, the concentration of virus in stool specimens from case-patient 7 dropped by 20-fold or more. Of note, this case-patient reported moderate diarrhea from days 2 to 12 of illness. Case-patient 6 had only mild diarrhea during the first 4 days of illness, and the amount of virus in his stool sample that was collected on day 19 (i.e., 2 weeks after the resolution of diarrhea) was approximately 800-fold lower than the amount in the day-14 stool sample of case-patient 7 and approximately 50-fold lower than that found in subsequent specimens from case-patient 7. No evidence was found that the virus mutated in case-patient 7 during the infection: genomic sequences of S and N genes of SARS-CoV from all positive stool specimens of this case-patient were identical.

No viral RNA was detected by RT-PCR in any of the eight urine specimens collected from the seven case-patients. SARS-CoV antibody was first found as early as days 10 and 11 after illness onset in three of seven case-patients. Adequate specimens were not available to characterize the time of first detectable SARS-CoV antibody in the remaining four case-patients (Table 3).

Household Transmission

Ten household contacts of five of the seven SARS case-patients were enrolled. Case-patient 1 had four household contacts, case-patients 3 and 4 had one such contact each, and case-patients 6 and 7 had two household contacts each. Of the 10 household contacts, 4 were female and 6 were male, 2 were smokers, and 2 reported previous history of respiratory problems (sarcoidosis in household contact 5 and pulmonary embolus in household contact 7).

Household contact 1 (who was also case-patient 2) was the only such contact who tested positive for SARS-CoV antibody. The remaining nine household contacts were negative for SARS-CoV antibody in specimens collected >28 days after their initial exposure to a case-patient. The infected household contact was the wife of confirmed SARS case-patient 1. The couple had visited Hong Kong together in early March 2003 and stayed at Hotel M, which was subsequently linked to the initial spread of SARS (16), where they had multiple opportunities for exposure. Case-patient 1 became ill 7 days after returning to the United States from Hong Kong. Symptoms developed in household contact 1 some 13 days after returning to the United States and 6 days after onset of illness in her husband. SARS-like symptoms did not develop in any of the three other household contacts of case-patient 1, nor did any have laboratory evidence of SARS-CoV infection. The analysis of household exposures and protective measures in this household indicated that household contact 1 had more frequent unprotected contact with the index patient compared with three other household contacts (Table 4).

The remaining six uninfected household contacts reported close contact (e.g., contact within 3 feet and unprotected skin-to-skin contact) with case-patients. The exposure of four household contacts of two case-patients with stool specimens positive for SARS-CoV was limited by isolation of the case-patients in a separate room with a private bathroom during the first week of illness. Both case-patients also wore surgical masks during this period, as did three of their four household contacts. Case-patient 7 was hospitalized from day 11 to day 18 of illness, the period during which the highest amounts of virus were detected in his stool, and continued to be positive for SARS-CoV in stool after discharge. Neither case-patient 7 nor his two household contacts wore surgical masks after being discharged from the hospital. Case-patient 6, who was never hospitalized, had low-level shedding of SARS-CoV in stool on day 19, but no virus was subsequently found in his stool specimens. One of his two household contacts wore a mask until 10 days after the resolution of fever in the case-patient.