Personal view of SARS: confusing definition, confusing diagnoses.

Sir

The diagnosis of severe acute respiratory syndrome (SARS) is based on a clinical definition. Overdiagnosis or underdiagnosis could happen, although to estimate the extent of this occurrence is difficult.

Overdiagnosis might lead to anxiety and fear associated with confinement and isolation, especially in children. In Hong Kong's public hospitals, health-workers are unable to provide a room for every patient with suspected or probable SARS. Patients diagnosed with SARS may or may not have the SARS virus (es), but they are at risk of contracting the infection if they are grouped with infected patients. Over-diagnosis also leads to inconvenience in the workplace or at school. The case of a young girl who developed fever during her visit to Taiwan is a good example. She had no SARS contact and chest radiograph was normal. However, the Taiwan government labelled her as a suspected SARS case, and she was transported back to Hong Kong where she was declared free from the infection, but the experience for her was not a memorable one.

Conversely, underdiagnosis of SARS could lead to cases of the infection being unrecognised, with the potential for pathogen to spread in the community. Imprecise definition therefore has serious public-health consequences.

WHO has done little to alleviate the confusion surrounding the terminology. SARS is an ambiguous term. The clinical features of many patients are neither severe nor respiratory in nature. SARS was initially labelled as atypical pneumonia with the outbreak in China, and this definition also caused much confusion with the so-called typical atypical pneumonia due to mycoplasma and chlamydia. Also, the acronym SARS is closely similar to that of ARDS for acute respiratory distress syndrome, but has a totally different meaning. Indeed, SARS typically kills patients not via the virus but by complications of ARDS.

WHO defines SARS as either suspected or probable. The case definition of suspected SARS is (1) fever, (2) respiratory symptoms including cough and difficulty breathing, and (3) close contact with people with SARS or a history of travel to an epidemic area. Probable SARS is a suspected case with radiographic evidence of pneumonia or respiratory distress syndrome. Thus, diagnosis of the cause of SARS was not required in these clinical definitions. However, to not obtain chest radiographs for a patient suspected as having SARS is not practical, neither is it sensible to label a patient as having suspected SARS in the absence of change on chest radiograph. Therefore, to differentiate suspected SARS from probable SARS on the basis of radiographic changes alone is not logical. Furthermore, any child in Hong Kong who has a cold with fever and cough would be diagnosed as suspected SARS by WHO definition. WHO has now revised their definition of a probable case of SARS to include a suspected case of SARS that is positive for SARS coronavirus.

In Hong Kong, the infection can be diagnosed by several definitions, official or otherwise. Suspected SARS, for instance, is now defined as SARS infection that is serious enough that ribavirin is started. The modification is a practical one and partly avoids Overdiagnosis of mild febrile symptoms that are probably not SARS. This modification, however, could lead to underdiagnosis of otherwise mild SARS cases. For example, a teenage girl was identified with symptoms of SARS and chest radiographic changes in accordance with the infection, but no definite history of contact was reported. She was started on ribavirin. Coronavirus was subsequently detected in her throat gargle and stool specimens by reverse transcriptase-PCR. She was labelled as suspected SARS in the absence of contact history. However, WHO would probably judge this case probable SARS because she resides in an epidemic area. If definite contact history is needed, this case is one of coronavirus pneumonia, not SARS at all. We have also seen other cases of SARS without fever and those with diarrhoea but no pneumonia.

To confuse matters further, the Chinese translation of probable SARS is confirmed SARS. In the Chinese media, confirmed SARS really means a person has been confirmed to have probable SARS. The local media must be reminded that SARS is really not the same as atypical pneumonia but rather a subset of this disease. For example, people can have atypical pneumonia but not SARS.

In view of the above confusions, we propose the term epidemic viral pneumonia (EVP) to replace suspected and probable SARS, and the classification shown in the table , which might be useful for index surveillance and in epidemiological and prognostication studies. The classification is not intended for triage of patients because any radiographic abnormality might not be noted at presentation. Virological results might also be available a few days after presentation. When we applied this classification to ten children previously reported, six could be grouped under EVP [C+, Coronavirus+] and four under EVP [C+, V-]. The teenage girl mentioned above would be classified as EVP [C-, Coronavirus+]. EVP [C-, V-] represents an overdiagnosed group of patients with various typical and atypical pneumonitis syndromes.

Table

Proposed classification system.

Classification	Definition
EVP [C+, V+]	EVP with positive contact history and virus identified
EVP [C+, V–]	EVP with positive contact history but no virus identified
EVP [C–, V+]	EVP with negative contact history but virus identified
EVP [C–, V–]	EVP with negative contact history and no virus identified

Our classification also helps to guide health-workers on patient's management. Newly admitted patients with persistent fever and pneumonia should be isolated, preferably in a single room, and be eventually classified into one of the four forms of EVP. Patients with EVP [C+, V+], EVP [C+, V-], and EVP [C-, V+] need to be isolated for at least 14 days, whereas those with EVP [C-, V-] could be discharged once their symptoms improve.