Article Text

Download PDFPDF

Diagnosis of Wilson's disease: an experience over three decades
  1. G J Brewer1
  1. 1Department of Human Genetics, University of Michigan Medical School, Buhl 4909, Ann Arbor, MI 48109-0618, USA brewergi{at}
  1. P J Gow2,
  2. R A Smallwood2,
  3. P W Angus2,
  4. R B Sewell2,
  5. A L Smith3,
  6. A J Wall4
  1. 2Department of Gastroenterology and Liver Transplantation, Austin and Repatriation Medical Centre, Heidelberg, 3084 Victoria, Australia Paul.Gow{at}
  2. 3Royal Children's Hospital, Parkville, 3052 Victoria, Australia
  3. 4Royal Melbourne Hospital, Parkville, 3052 Victoria, Australia

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

The paper by Gow et al (Gut 2000;46:415–19) discussed the diagnosis of Wilson's disease in 30 patients presenting to two different clinical facilities over 28 years (1971–1998). Because a paper of this type is likely to be viewed as an authoritative guide, it is important that the information be valid. For that reason, I call attention to the following significant errors in the paper.

The authors report urine copper values of 5, 4, 7, 4, 5, 2, and 2 μg per 24 hours in seven patients in their table 1. These data cannot possibly be valid. The normal range for urine copper is 20–50 μg per 24 hours, and these patients are far below the lower limit of normal. In performing several thousand 24 hour urine copper tests on patients and normal subjects in our own laboratory, I have never seen one below 10 μg per 24 hours, except in copper deficiency. The patients in table 1 have Wilson's disease, the opposite of copper deficiency, making the data even more unbelievable. I also do not believe the data for two additional symptomatic patients in their table 1 who are reported to have urine copper values of 55 and 44 μg per 24 hours. Urine copper values in untreated symptomatic Wilson's disease patients are invariably over 100 μg. That is our experience in all of 88 newly diagnosed neurologically presenting patients (urine copper range 107–1593 μg/24 hours) and in all of 18 newly diagnosed patients with hepatic presentation (urine copper range 106–1880 μg/24 hours). Not all liver disease patients with urine copper values over 100 μg will have Wilson's disease but all untreated patients with Wilson's disease with clinically presenting liver disease will have a value over 100 μg. If a patient with liver disease does not have a value that high, then look to another diagnosis, or to a laboratory error. No doubt the latter is the case here because the values in seven of the patients are not biologically reasonable. One caveat: if the patient has been treated with a chelating agent, even briefly, and then the drug stopped, there is often a rebound period when urine copper will drop below 100 μg.

I also question the diagnosis of Wilson's disease in two of the patients in this series. Patient Nos 20 and 22 had liver disease but were negative for Kayser-Fleischer rings, had a non-elevated urine copper, and normal caeruloplasmin levels. The diagnosis was supposedly made in patient No 20 by elevated liver copper. Non-Wilsonian chronic liver disease can also elevate hepatic copper.1,2 The diagnosis in this patient is uncertain, and disproved if urine copper is truly low. Patient No 22 also had elevated liver copper and a “positive radiolabelled copper plasma clearance test.” Again, hepatic copper is not proof of the diagnosis. The abnormal radiocopper test is supportive but Wilson's heterozygotes, comprising about 1% of the population, are often falsely positive.3 Urine copper of 44 μg, if valid, would rule out the diagnosis.

In summary, if the data in the current paper were to be believed (nine of 22 patients with low or normal copper values), it would inappropriately denigrate the great usefulness of urine copper assay in the diagnosis of symptomatic Wilson's disease. Secondly, the liver copper value is not an absolute in the presence of chronic liver disease. Sometimes a therapeutic trial is the only way to come to a decision in these patients.


Author's reply

Dr Brewer in his letter raises two points regarding our recent paper (Gut 2000;46:415–9). Most importantly, he questions the diagnosis of Wilson's disease in two patients. Both of these patients were very unusual in that they were both older than 40 years at diagnosis, had caeruloplasmin levels within the normal range, no Kayser-Fleischer rings, and low urinary copper excretion. The diagnosis in these two interesting patients is discussed in the paper. To summarise the discussion: in both patients the diagnosis was based on the findings of markedly elevated liver copper levels (634 and 1037 μg/g dry weight liver, normal liver copper 15–60 μg/g). In addition, the first patient had a family history of cirrhosis and the second had a positive radiolabelled copper plasma clearance test. In this situation the primary differential diagnosis relates to elevated liver copper from chronic cholestatic liver disease. In both of these patients, chronic cholestatic liver disease was excluded by the absence of cholestatic biochemistry plus a lack of cholestatic findings on histological examination of the explanted liver. The important point to emphasise with these cases is the lack of sensitivity of caeruloplasmin, urinary copper excretion, and Kayser-Fleischer rings in the diagnosis of Wilson's disease.

Dr Brewer's second point relates to reported low levels of urinary copper excretion in seven of our patients. In these patients the values were reported as μmol/24 hours when they should have been converted to μg/24 hours. The true levels for patient Nos 6, 7, 8, 11, 12, 17, and 20 are 312, 117, 437, 1250, 762, 125, and 100 μg/24 hours, respectively.

Dr Brewer also questioned the urinary copper excretion values in patient Nos 16 and 22. In both of these cases the data were rechecked and the reported values were found to be correct.

Linked Articles