ARTICLE Severe Acute Respiratory Syndrome Associated Coronavirus Infection in Toronto Children A Second Look AriBitnun MD MSc a StanleyRead MD PhD a RaymondTellier MD MSc b MartinPetric PhD c SusanE Richardson MD b a Division of Infectious Diseases Department of Pediatrics and b Department of Pediatric Laboratory Medicine Hospital for Sick Children University of Toronto Toronto Ontario Canada c British Columbia Center for Disease Control Vancouver British Columbia Canada The authors have indicated they have no financial relationships relevant to this article to disclose What sKnownonThisSubject TheclinicalcourseofSARS CoVinfectiontendstobemilderinyoungchildrencompared with teenagers and adults Young children also seem to be less infectious than older persons Most of the pediatric experience with SARS CoV infection is from Southeast Asia WhatThisStudyAdds ThisstudycohortrepresentstheonlyNorthAmericanexperiencewithSARS CoVinfec tion in children The importance of epidemiology in case detection is a key finding Additional evidence for the mild nature of SARS CoV infection in childrenH1102112 years of age is provided ABSTRACT OBJECTIVES During the severe acute respiratory syndrome outbreak of 2003 there was an impetus to provide clinical information to the medical community in a timely manner Accordingly a preliminary report of our experience of suspected severe acute respiratory syndrome associated coronavirus infections in children was pub lished without microbiological findings This report provides an update on pediatric severe acute respiratory syndrome associated coronavirus infections in Toronto Ontario Canada that includes microbiological findings METHODS All of the children admitted to the Hospital for Sick Children between March 14 and June 15 2003 with suspect severe acute respiratory syndrome associated coronavirus infection were included A proven case was defined as one that fulfilled the clinical criteria for suspect severe acute respiratory syndrome associated coro navirus infection and demonstrated a serologic response to severe acute respiratory syndrome associated coronavirus Serology results from a neutralizing antibody assay were considered positive if the sera inhibited the development of a severe acute respiratory syndrome associated coronavirus specific cytopathic effect at a dilution of H113501 8 RESULTS Neutralizing antibody to severe acute respiratory syndrome associated coro navirus was demonstrated in 8 of 25 children admitted with suspect severe acute respiratory syndrome associated coronavirus infection In 3 of these 8 children severe acute respiratory syndrome associated coronavirus was also detected by reverse transcription polymerase chain reaction in the stool All 8 had documented exposure to H113501 severe acute respiratory syndrome associated coronavirus infected adults residing in the same household Exposure that was limited to visiting a Toronto hospital at which severe acute respiratory syndrome associated coronavirus infected patients were admitted or travel from a country in which severe acute respiratory syndrome had been reported did not result in documented infection in any of our cases On the basis of our clinical case definition 6 of 8 microbiologically confirmed case had been classified as having probable severe acute respiratory syndrome associ ated coronavirus infection Clinical disease was mild nonspecific and self limited and was indistinguishable from that reported with other common respiratory viruses CONCLUSIONS The factor most strongly associated with severe acute respiratory syndrome associated coronavirus infection in Toronto children was a history of close contact with an adult severe acute respiratory syndrome associated coronavirus case This serves to reinforce the importance of routinely obtaining a thorough epidemiologic travel and exposure history for all subjects with suspected infectious diseases Pediatrics 2009 123 97 101 H ORSESHOE BATS IN southern China have been shown recently to be the natural reservoir for coronaviruses phylogenetically closely related to the severe acute respiratory syndrome associated coronavirus SARS CoV 1 2 It has also been shown that SARS CoV is capable of infecting a wide range of wild and domestic mammals 3 4 Given the frequent close contact of humans with animals potentially infected with SARS CoV or related viruses in www pediatrics org cgi doi 10 1542 peds 2007 3745 doi 10 1542 peds 2007 3745 KeyWords severe acute respiratory syndrome SARS reverse transcription polymerase chain reaction probable SARS suspect SARS Abbreviations SARS severe acute respiratory syndrome CoV associated coronavirus RT PCR reverse transcription polymerase chain reaction Accepted for publication Mar 25 2008 Address correspondence to Ari Bitnun MD MSc FRCPC Hospital for Sick Children Division of Infectious Diseases 555 University Ave Toronto Ontario Canada M5G 1X8 E mail ari bitnun sickkids ca PEDIATRICS ISSNNumbers Print 0031 4005 Online 1098 4275 Copyright 2009bythe AmericanAcademyofPediatrics PEDIATRICS Volume 123 Number 1 January 2009 97 at Dahlgren Medical Library on March 9 2015pediatrics aappublications orgDownloaded from the wet markets of southern China and to a lesser extent in the wild it is quite possible that SARS CoV or a related virus could be reintroduced into the human pop ulation in the future We described previously the clinical laboratory and radiographic features of children with suspect or proba ble SARS CoV infection admitted to the Hospital for Sick Children during the SARS outbreak that occurred in Toronto Ontario Canada between March and June of 2003 5 In that report children were categorized on the basis of potential SARS exposure clinical features and the presence or absence of an alternative microbiological diagnosis In this report we provide an update on pedi atric SARS CoV infection in Toronto based on microbi ological diagnosis METHODS All of the children admitted to the Hospital for Sick Children between March 14 and June 15 2003 with suspect SARS were included in this report The clinical case definitions and management protocol used during the SARS outbreak were described in our previous re port 5 Briefly children were classified as having suspect SARS CoV infection if they had a history of possible SARS exposure and fever H1102238 C A probable case was defined as suspect SARS CoV infection plus severe pro gressive respiratory illness and or chest radiograph find ings indicative of lower respiratory tract disease without microbiological evidence implicating another causative agent Possible SARS CoV infection was defined as a suspect case without the aforementioned clinical and radiographic criteria for probable SARS CoV infection without microbiological evidence implicating another causative agent The other etiology category was re served for those in whom another causative agent was identified In the current report a proven case of SARS CoV infection was defined as one that fulfilled the clin ical criteria for suspect SARS CoV infection and demon strated neutralizing antibody to SARS CoV in acute and convalescent sera Those fulfilling the clinical case defi nition for suspect SARS CoV infection who were sero negative in their acute and convalescent sera were ac cordingly defined as non SARS CoV case subjects SARS CoV serology was performed by a neutralizing antibody assay in a dedicated biosafety level III facility at the British Columbia Center for Disease Control 6 Sera were tested for their ability to neutralize the virus from dilutions of 1 8 to 1 4096 Acute and or convalescent sera that exhibited viral neutralization at a dilution of H113501 8 were deemed positive for neutralizing antibody Reverse transcription polymerase chain reaction RT PCR on stool and nasopharyngeal specimens was per formed as follows A 10 suspension of the stool in double distilled water was clarified by centrifugation and the supernatant fraction processed for nucleic acid extraction The specimens containing the nasopharyn geal swabs were agitated on a vortex mixer and the transport medium subjected to nucleic acid extraction RNA was extracted from the specimens by a semiauto mated method using the MagaZorb RNA kit Cortex Biochem Inc San Leandro CA with the KingFisher instrument Thermo Electron Corp Waltham MA Each sample was tested with 2 different assays an in house assay for coronaviruses 7 and a commercial assay the Real Art HPA Coronavirus RT PCR Artus GmbH Hamburg Germany RESULTS Twenty five children were admitted to the Hospital for Sick Children between March 14 and June 15 2003 with a diagnosis of suspect SARS CoV infection Neu tralizing antibody to the SARS CoV was demonstrated in 8 of these children SARS CoV RNA was detected by RT PCR in the stool of 3 children by both RT PCR meth ods all 3 also had a neutralizing antibody to the virus The timing of viral detection in stool ranged from day 5 to day 7 of illness Thirteen nasopharyngeal swabs and 8 nasopharyngeal aspirates on 15 patients were tested and found to be negative by the real time RT PCR assay In 1 patient with serologically confirmed SARS CoV infec tion coinfection with adenovirus was demonstrated by the detection of adenovirus antigen in a nasopharyngeal swab In 9 of the 17 children with negative SARS CoV serology results an alternate etiology was identified including influenza A virus n H11005 3 parainfluenza 3 virus n H11005 1 respiratory syncytial virus n H11005 1 ade novirus n H11005 1 varicella zoster virus n H11005 1 rotavirus n H11005 1 and Streptococcus pneumoniae n H11005 1 According to the clinical diagnostic criteria used during the SARS outbreak 6 serologically confirmed cases had been clas sified as probable SARS CoV infections 1 as a possible SARS CoV infection and 1 as having another etiology The clinical laboratory and radiographic features of serologically confirmed SARS CoV cases are shown in Table 1 and a comparison of clinical and laboratory features of SARS CoV and non SARS CoV cases is given in Table 2 Although the clinical manifestations of those with and without SARS CoV infection were indistin guishable respiratory symptoms were observed in only 3 children 37 5 in the former group However chest radiograph abnormalities consisting predominantly of focal infiltrates were noted in 87 5 7 of 8 of SAR CoV cases compared with 29 5 of 17 of non SARS CoV cases Fisher s exact test P H11005 01 Diarrhea oc curred in 2 children with SARS CoV infection Importantly with the exception of one 17 5 year old girl who required supplemental oxygen by nasal prongs the illness was so mild that hospital admission and or supportive therapy of any kind would not have been needed had SARS not been a consideration 5 Hematologic abnormalities were noted in the major ity of SARS CoV infections At the time of admission leukopenia was observed in 50 4 of 8 of those with compared with 6 1 of 17 of those without SARS CoV infection Fisher s exact test PH11005 02 Lymphopenia was also more common in the SARS CoV group at admission 50 vs 18 Fisher s exact test P H11005 16 whereas neutropenia was relatively uncommon occurring in only 1 child in each group During the course of hospi talization leukopenia was seen in 75 0 6 of 8 and neutropenia in 62 5 5 of 8 of SARS CoV infections compared with 18 0 3 of 17 for both parameters in 98 BITNUN et al at Dahlgren Medical Library on March 9 2015pediatrics aappublications orgDownloaded from the non SARS CoV group Fisher s exact test P H11005 01 and P H11005 06 respectively In SARS CoV infected chil dren the nadir lymphocyte count occurred within 24 hours of hospitalization in 62 5 5 of 8 of cases whereas the neutrophil nadir was seen between days 4 and 6 of hospitalization in a similar proportion Alanine aminotransferase was slightly higher in the SARS CoV group Kruskal Wallis test P H11005 01 There were no TABLE 2 ClinicalandLaboratoryFeaturesofChildrenWithandWithoutSerologicallyConfirmed SARS CoVInfection Parameter SARS CoV nH110058 Non SARS CoV nH1100517 P a Age y 5 5 2 1 11 5 2 0 1 2 3 0 14 Gender female 50 65 67 Temperature C 38 4 38 2 40 2 39 3 39 0 39 7 75 Cough 37 5 82 0 06 Coryza 12 5 41 0 20 Diarrhea 25 0 24 0 99 CXR abnormalities 87 5 29 0 01 Admission leukocyte count H1100310 9 L 6 0 4 1 11 6 9 3 6 3 11 1 17 Leukocyte count nadir H1100310 9 L 4 0 2 9 5 5 8 2 5 6 9 8 01 Admission neutrophil count H1100310 9 L 2 5 2 2 4 7 5 24 3 2 8 4 07 Neutrophil count nadir H1100310 9 L 1 2 0 6 1 8 3 3 1 9 5 5 01 Admission lymphocyte count H1100310 9 L 1 6 1 1 4 8 2 5 1 5 3 5 46 Lymphocyte count nadir H1100310 9 L 1 3 1 1 3 3 2 1 1 5 3 5 31 Admission platelet count H1100310 9 L 312 222 345 254 216 315 52 Platelet count nadir H1100310 9 L 221 158 264 243 172 294 41 ALT U L 34 27 57 16 4 21 01 AST U L 42 38 79 41 29 53 35 CPK U L 142 101 200 78 57 160 07 LDH U L 1581 1012 2150 828 733 850 12 Data are reported as medians interquartile ranges for continuous variables and proportions for dichotomous variables a Data were calculated by the Fisher s exact test for dichotomous variables and Kruskal Wallis test for continuous variables TABLE 1 ClinicalandLaboratoryFeaturesofSARS CoV InfectedPatients Case Age y Gender T max C Clinical Manifestations Chest Radiograph Findings Laboratory Abnormalities SARS CoV Neutralization Inhibition Titer Acute Convalescent a 1 17 5 F 40 1 Cough dyspnea hypoxemia bilateral crackles Dense RML and LLL infiltrate Leukopenia 2 70H1100310 9 L lymphopenia 0 81H1100310 9 L thrombocytopenia 130H1100310 9 L elevated AST ALT 236 U L 187 U L elevated CPK 457 U L elevated LDH H110222150 U L H110211 8 1 32 2 12 F 38 1 Cough diarrhea LLL infiltrate None H110211 8 1 128 3 11 M 38 5 Headache chills RLL infiltrate Lymphopenia 1 39H1100310 9 L H110211 8 1 256 4 6 M 38 2 Sore throat vomiting Patchy lower lobe infiltrate peribronchial thickening Lymphopenia 1 30H1100310 9 L neutropenia 0 33 x10 9 L elevated AST 61 U L elevated LDH 1012 U L 1 64 1 128 5 5 F 40 3 Cough Normal Leukopenia 3 10H1100310 9 L neutropenia 1 00H1100310 9 L lymphopenia 1 30H1100310 9 L ND 1 64 6 2 4 M 38 2 Lethargy RLL infiltrate Leukopenia 2 4H1100310 9 L lymphopenia 0 74H1100310 9 L neutropenia 0 63H1100310 9 L H110211 8 1 512 7 1 75 F 39 0 Diarrhea lethargy Multifocal perihilar and lower lobe patchy infiltrates Neutropenia 1 4H1100310 9 L 1 8 1 64 8 0 4 M 39 0 Rhinorrhea Patchy RUL and RLL infiltrates Neutropenia 0 65H1100310 9 L elevated AST ALT 97 U L 74 U L 1 64 1 64 F indicates female M male RUL right upper lobe RLL right lower lobe RML right middle lobe LLL left lower lobe AST aspartate aminotransferase ALT alanine aminotransferase CPK creatine kinase LDH lactate dehydrogenase T max maximum temperature a ND indicates not done acute sample was not available for testing PEDIATRICS Volume 123 Number 1 January 2009 99 at Dahlgren Medical Library on March 9 2015pediatrics aappublications orgDownloaded from differences between the 2 groups with respect to platelet count aspartate aminotransferase creatine kinase or lactate dehydrogenase All 8 of the children with serologically confirmed SARS CoV infection had documented exposure to H113501 adult suspect or probable SARS case who resided in the same household Table 3 By contrast none of the 14 children with travel related or hospital related expo sures were found to have SARS CoV infection Docu mented SARS exposure was thus highly predictive of SARS CoV infection Fisher s exact test PH11021 0002 The sensitivity specificity positive predictive value and neg ative predictive value of documented direct exposure to a suspect SARS case were 100 82 73 and 100 respectively All 8 of the children confirmed to have been infected with SARS CoV were seen in follow up within 1 to 2 month of discharge None reported any persistent or new symptoms including fatigue exercise intolerance dyspnea or wheezing All of the caregivers and children felt that their health was fully restored Four of the older children including the 17 year old who required oxy gen while hospitalized were seen 12 to 15 months after discharge and reported to be in good health All had normal chest radiographs at this time Pulmonary func tion testing including prebronchodilator and postbron chodilator responses was performed in the aforemen tioned 17 year old and found to be normal DISCUSSION This series provides additional evidence that SARS CoV infection in children is relatively mild and nonspecific and further supports our decision at the time of the outbreak to use less stringent clinical diagnostic criteria than those proposed by the World Health Organization Although all children with microbiologically confirmed SARS CoV infection had fever only 3 37 5 of 8 had cough and only 1 had dyspnea Chest radiograph ab normalities consisting of minor nonspecific alveolar in filtrates were evident in 7 of 8 patients in some despite the absence of respiratory symptoms As has been ob served by others 8 9 the most reliable clue to the diagnosis was an epidemiological link to a suspected or confirmed SARS case In our setting all of the children with sero logically confirmed SARS CoV infection had docu mented exposure to at least 1 suspected adult case who resided in the same household Although the degree to which this would hold true in the event of a future SARS CoV outbreak is difficult to predict it is likely that detection of potential cases on the basis of an epidemi ological link early in the course of an outbreak would allow for rapid and efficient containment of the out break Based on our experience as well as that of oth ers 8 9 we would recommend that in the context of a SARS CoV outbreak any child with fever and an epide miological link particularly if H1102112 years of age and irrespective of respiratory symptoms be investigated for possible SARS CoV infection Hematologic abnormalities are relatively common among children with SARS CoV infection 8 14 In our se ries leukopenia at the time of admission and neutrope nia during the course of hospitalization were more com mon in SARS CoV infected subjects than in uninfected subjects although these observations are not likely to be sufficiently discriminatory for diagnostic purposes 9 The nadir lymphocyte count tended to occur early in the course of illness whereas neutropenia tended to develop later on between days 4 and 6 of hospitalization It is important to note that none of our children were treated with corticosteroids a fact that could at least partially explain the absence of progressive lymphopenia that was observed in other studies 8 15 It is also conceivable that in mildly affected patients such as ours progressive lym phopenia is less likely to develop The ability to rapidly and reliably confirm the diag nosis of SARS CoV infection remains elusive because of the relative insensitivity of direct detection methodolo gies such as RT PCR during the early phase of the illness unless lower respiratory tract samples are avail able for testing SARS CoV was detected in 47 7 of nasopharyngeal aspirates and in 38 6 of stool samples of serologically confirmed patients in the largest pediat ric series published to date 8 In most patients SARS CoV was detected in the nasopharynx during the first week of illness whereas for stool samples this tended to occur 7 days into the illness or beyond In our cohort SARS CoV was detected by RT PCR in the stool of 37 5 of subjects between days 5 and 7 of illness but in no child was it detected in nasopharyngeal samples It is possible that our failure to detect the virus in the nasopharynx was a consequence of us obtaining samples within 24 hours of admission a time when the viral burden in the naso pharynx may be quite low 16 The optimal timing for nasopharyngeal sampling for children with a mild and brief SARS CoV illness has not been established but it is likely that the viral burden in such children never reaches the levels observed in se verely affected adolescents and adu