[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
November 26, 2003

Temporal Patterns of Hepatic Dysfunction and Disease Severity in Patients With SARS

Author Affiliations

Letters Section Editor: Stephen J. Lurie, MD, PhD, Senior Editor.

JAMA. 2003;290(20):2663-2665. doi:10.1001/jama.290.20.2663

To the Editor: Mildly elevated levels of aminotransferase have been reported in 23% to 50% of patients with severe acute respiratory syndrome (SARS). The clinical and pathological significance of this finding, however, remains unknown.13 We investigated whether liver dysfunction might reflect disease activity and have a temporal relationship with other clinical parameters in patients with SARS.


We evaluated clinical records, serial blood test results, and chest radiographs of 54 adult patients with SARS (diagnosed according to US Centers for Disease Control and Prevention criteria1,4) but without chronic hepatitis B or C infection, who were admitted to 2 Hong Kong regional hospitals between March 8 and April 17, 2003. Charts of randomly selected age- and sex-matched patients with non-SARS community-acquired pneumonia from Queen Mary Hospital in the same time period were also examined. All patients with SARS received combination therapy comprising a corticosteroid and either intravenous or oral ribavirin, beginning a mean of of 3.0 (SD, 1.5) days from date of admission. All patients were also treated with intravenous cefepime (2 g 3 times daily) and with either oral clarithromycin (500 mg twice daily) or azithromycin (500 mg once daily), beginning at admission and continuing throughout their treatment period, as standard protocol for treatment of severe community-acquired pneumonia in our institution.

Liver dysfunction was defined as alanine aminotransferase (ALT) level greater than the upper limit of normal. The time from admission to peak ALT level was determined for each patient. Serial chest radiographs of the 54 patients with SARS were quantitatively evaluated to provide a daily "chest x-ray (CXR) score," which reflected the degree of lung involvement. We recorded the baseline CXR score, the maximal CXR score (worst radiographic disease severity) as well as the number of days between admission and the day on which the maximal CXR score was noted.We also recorded the posttreatment CXR score, obtained either on the last day of assessment or after at least 96 hours of defervescence together with radiographic evidence of improvement in lung consolidation.

Data were analyzed with the t test, the Spearman rho correlation, or the Pearson correlation coefficient as appropriate. A P value less than .05 was considered statistically significant. Our study was approved by a local institutional review board.


Compared with patients with non-SARS community-acquired pneumonia, patients with SARS had significantly lower initial total leukocyte and lymphocyte counts, as well as globulin levels, and also had significantly higher initial ALT levels. (Table 1 and Table 2). Liver dysfunction was found in 29.6% and 75.9% of patients at presentation (before ribavirin treatment) and during treatment, respectively (P<.001 by χ2 test).

Table 1. Demographic Data for Patients With SARS and for Age- and Sex-Matched Controls Without SARS
Image description not available.
Table 2. Laboratory Results for Patients With SARS and for Age- and Sex-Matched Controls Without SARS
Image description not available.

Among patients with SARS, levels of albumin, globulin, bilirubin, aspartate aminotransferase, and ALT all changed significantly during treatment (Table 2). Time to maximal CXR score (mean [SD], 7.2 [4.0] days) correlated with time to peak ALT level (mean [SD], 11.6 [6.6] days; r = 0.51, P<.001). Albumin levels on admission and during time to peak ALT level both had a significant negative correlation with maximal CXR score (r = −0.39, P = .008; and r = −0.52, P<.001, respectively). Bilirubin levels during time to peak ALT level correlated with maximal CXR score (r = 0.35, P = .02). Bilirubin and albumin levels during time to peak ALT level correlated with the posttreatment CXR scores (r = 0.37, P = .008; and r = −0.36, P = .01, respectively).

At the posttreatment assessment, mean levels of albumin and aspartate aminotransferase had returned to normal, while mean levels of ALT and bilirubin were still elevated (Table 2). The mean cumulative dosage and duration of ribavirin treatment were similar between those who developed liver dysfunction and those who did not (data not shown).


The pathological resemblance of SARS to diffuse alveolar injury, as well as its radiologic resemblance to bronchiolitis obliterans organizing pneumonia, suggest that the pathogenesis of SARS may involve an immunological component.13 Given the clinical and radiological responsiveness of SARS pneumonia to corticosteroid therapy, our findings of a temporal relationship between liver dysfunction and pneumonic severity suggest that liver dysfunction in SARS might reflect an immunological injury to the liver. The temporal relationship between maximal liver dysfunction and pneumonic severity also suggest that a common pathogenic mechanism is in action for both organs. We acknowledge that ribavirin may have also played a role in this relationship. We think that a drug-induced hepatitis is a less likely explanation for our findings, however, as most patients had evidence of liver dysfunction prior to therapy. Furthermore, this would not necessarily explain the correlations that we found between levels of albumin and bilirubin during time to peak ALT level and the severity of radiographic disease. While coronavirus causes fulminant hepatitis in mice5,6 and thus it is possible that coronavirus could directly damage the liver, this is less likely to be the case in light of the temporal relationship to the steroid-responsiveness of the lung disease.

Our data, therefore, suggest that liver dysfunction is a distinct abnormality in patients with SARS, and further suggest a temporal relationship between clinical disease severity and liver dysfunction in SARS. Since there are no specific markers for this disease, liver function monitoring may be a useful adjunct to serial radiographic assessment in the clinical assessment of patients with SARS.

Acknowledgment: We would like to thank the doctors, nurses, and medical personnel of Queen Mary and Queen Elizabeth Hospitals for their dedication and care of SARS patients. We would also like to thank Kennis To, Fiona Fung, Teresa Tong, June Sun, Colin Ko, and Christina Yan for data management. Drs W.-M. Wong and J. C. Ho contributed equally to this work.

Tsang KW, Ho PL, Ooi GC.  et al.  A cluster of cases of severe acute respiratory syndrome in Hong Kong.  N Engl J Med.2003;348:1977-1985.PubMed
Lee N, Hui D, Wu A.  et al.  A major outbreak of severe acute respiratory syndrome in Hong Kong.  N Engl J Med.2003;348:1986-1994.PubMed
Peiris JSM, Lai ST, Poon LLM.  et al.  Coronavirus as a possible cause of severe acute respiratory syndrome.  Lancet.2003;361:1319-1325.PubMed
US Centers for Disease Control and Prevention (CDC).  Updated interim U.S. case definition for severe acute respiratory syndrome (SARS). Atlanta, Ga: CDC; 2003. Available at: http://www.cdc.gov/ncidod/sars/casedefinition.htm. Accessibility verified September 30, 2003.
Ding JW, Ning Q, Liu MF.  et al.  Fulminant hepatic failure in murine hepatitis virus strain 3 infection: tissue-specific expression of a novel fgl2 prothrombinase.  J Virol.1997;71:9223-9230.PubMed
Kyuwa S, Shibata S, Tagawa Y, Iwakura Y, Machii K, Urano T. Acute hepatic failure in IFN-gamma-deficient BALB/c mice after murine coronavirus infection.  Virus Res.2002;83:169-177.PubMed