Editor—We appreciated the report by O'Beirne and Cairns concerning a patient with liver test dysfunction in the setting of treatment with the COX 2 inhibitor celecoxib.1 We are, however, concerned at their use of “cholestatic hepatitis” as the most appropriate description of the pattern of liver test abnormality observed. The patient they described had a maximal aspartate transaminase concentration of 1650 IU/l (reference range 10-40 IU/l), a maximal alkaline phosphatase concentration of 232 IU/l (25-115 IU/l), and peak total serum bilirubin of 123 μmol/l (5-20 μmol/l). In broad terms, two categories of drug associated liver injury are encountered commonly, namely cholestatic and hepatocellular.2 Cholestatic injury has been defined further as occurring when the peak transaminase concentration is less than eight times the upper limit of normal, and the corresponding ALP is greater than threefold normal, whereas hepatocellular injury has been defined as being present when the peak transaminase concentration is greater than eight times the upper limit of normal and the concomitant alkaline phosphatase concentration is less than threefold normal. A mixed pattern of injury, showing features of both, may also be found.3 According to these criteria the patient of O'Beirne and Cairns had evidence of hepatocellular injury primarily, rather than a mixed pattern as the term “cholestatic hepatitis” suggests. Liver biopsy might have helped to emphasise this distinction. The article by Maddrey et al referred to in their report was misquoted4: where the term alkaline phosphatase was used, it should have read alanine aminotransferase. In that study only 0.4% and 0.3% respectively of 6376 patients treated with celecoxib had maximal alanine aminotransferase and aspartate transaminase concentrations greater than or equal to three times the upper limit of normal. None of these transaminase elevations was greater than eight times the upper limit of normal, in contrast to that found in the patient of O'Beirne and Cairns (about 41 times upper normal limit). Therefore, although we disagree with the view that this patient had cholestatic hepatitis on the basis of data quoted, the case does represent the first reported instance of severe hepatocellular liver dysfunction in association with celecoxib treatment.