Wilson's disease (WD), first described by Samuel Alexander Kinnier Wilson (KW) in 1912, is an autosomal recessive metabolic disorder resulting from mutations in the ATP7B gene (Bull 1993; Compston 2009). This secretory pathway involves both copper excretion into bile and its incorporation into apocaeruloplasmin for the synthesis of functional caeruloplasmin (Davis 1996). The disease develops as a consequence of the accumulation of copper in affected tissues.

The clinical presentation of WD can vary widely in terms of symptoms, signs at presentation and age of onset. The key clinical diagnostic features used to form the basis of the Leipzig criteria include liver disease, motor and neuropsychiatric disturbances, corneal Kayser-Fleischer (KF) rings and acute haemolysis (in association with acute liver failure) (Ferenci 2003). The original classical neurological presentations as described by KW, are characterised by a movement disorder in the setting of characteristic biochemical abnormalities and often contrast with the non-specific protean hepatic manifestations.

The diagnosis of WD is often delayed in both children and adults, which may potentially affect outcome and have implications for other family members. The diagnosis of WD depends on a combination of clinical, biochemical, histological and genetic testing and analysis. The Leipzig criteria ( Table 1) were established to help standardise diagnosis and management (Ferenci 2003). However, it should be emphasised that these criteria date from 2003, and many of these have not been formally evaluated; this review will examine the evidence behind biochemical testing for WD (Mak 2008).

The most commonly used initial diagnostic test for WD is caeruloplasmin, with a concentration of less than 0.2 g/L considered as the conventional diagnostic cut off (EASL 2012). However, the lower reference limit can vary with different assay types, patient age and may be reduced in other causes of chronic liver disease, copper deficiency or in protein-losing states. Caeruloplasmin was originally described as an acute phase protein with diverse functions (Hellman 2002). The initial protein produced is an inactive, unstable non-copper bound form, apocaeruloplasmin. Following the addition of copper by ATP7B, the functional more stable product holocaeruloplasmin is formed. The type of assay used has important implications for WD diagnosis. Immunoassays are commonly used for measuring caeruloplasmin, and measure both apo- and holocaeruloplasmin forms; however, caeruloplasmin oxidase-based methods only measure the holocaeruloplasmin form. Therefore, immunoassays may theoretically lead to an overestimate as compared with the caeruloplasmin oxidase-based method. Assay availability may also be an issue (Gnanou 2006; Walshe 2003).

Twenty-four hour urinary copper studies are often used as a follow up to abnormal caeruloplasmin testing. In the absence of renal impairment, urinary copper reflects the amount of non-caeruloplasmin bound copper. Cut-off values of more than 1.2 times the upper limit of normal (ULN) or more than 2 ULN have been suggested as indicating possible WD (EASL 2012). Use of such cut-off values is problematic, method dependent and up to 25% of people with WD may have urinary copper levels less than this, especially children (Nicastro 2010). Urinary copper excretion may also be increased in other causes of chronic liver disease (LaRusso 1976).

Hepatic copper accumulation is the hallmark and earliest manifestation of WD. Copper distribution within hepatic parenchyma may not be homogenous, may be susceptible to sampling error by biopsy and may be elevated in other liver disorders, particularly those involving cholestasis (Roberts 2008). Cut-off values have been suggested, but again these are method-dependent and as yet not fully validated (EASL 2012). Use of specific stains, such as rhodamine or orcein, reveal focal copper deposition in less than 10% as these only detect lysosomal deposition. Previous studies have suggested that a level of hepatic copper greater than 4 µmol/g is considered the best evidence for a diagnosis of WD; however, there is some evidence that such a threshold may need to be lower in order to increase sensitivity (Yang 2015).

In the original paper outlining the Leipzig criteria, there is no agreed clinical reference standard pathway for the diagnosis of WD, and hence this has not been documented (Ferenci 2003). There is a diagnostic algorithm in the EASL guidelines on WD, but again this is not a complete pathway that outlines test selection and stages of testing proceeding to diagnosis (EASL 2012). The EASL guidelines do comment that a combination of tests are required and that, reflecting the challenge of diagnosing WD, no single test is specific.

Alternative test(s)

This autosomal recessive disorder of copper transport is due to mutations in the ATP7B gene. The worldwide prevalence of WD has previously been cited as 1 in 30,000, with a carrier frequency of 1 in 90; however, these figures pre-date the discovery of the ATP7B gene and more recent work has cited a higher frequency of 1 in 7000 for genetic diagnosis (Coffey 2013). Following extensive linkage and positional cloning studies, the ATP7B gene was located on chromosome 13q14.3 (Bull 1993). The gene has 21 exons with more than 10000 base pairs. The molecular analysis of individuals and families affected by WD have demonstrated that, to date there are up to 500 disease-causing mutations (Coffey 2013).

The problem with the collation of such mutations and variants is the lack of controls tested in studies, which then inaccurately reported new variants. Recommendations of a minimum of 100 normal chromosomes from the same ethnic population to be tested are often not followed (Kenney 2007). Whilst most of the pathogenic mutations discovered to date are rare and only reported in single families, some of these are more common and account for large numbers of WD cases. Most affected individuals are compound heterozygotes. These mutations mainly affect the transmembrane region and largely consist of missense and stop mutations (Kenney 2007; Thomas 1995). However, strict genotype-phenotype remains unproven, even within families and therefore other genetic modifiers are believed to be at play (Czlonkowska 2009; Huster 2012).

In the presence of definite clinical or biochemical abnormalities, the identification of only one of the two disease-causing genes may be adequate to confirm diagnosis. However, if the significance of the initial identified mutation is doubtful, the second mutation should be identified (EASL 2012). It should be noted that in order to infer pathogenicity, a mutation must clearly be disease-causing and not just a common missense variant. Hence the importance of normal ethnic controls. Developments in next-generation sequencing may allow faster sequencing and better coverage; however, large numbers of variants of uncertain significance may be generated and relevant standardised functional methods to test these remain to be clearly established.

Exchangeable copper and its derived relative exchangeable copper (REC) have recently been proposed as a new biomarker for diagnosing WD. Specfically, REC has been shown to provide a high sensitivity and specificity for WD (El Balkhi 2011). The exchangeable copper corresponds to the labile fraction of copper bound mainly to albumin as well as free unbound copper. An increase in this fraction above normal is thought to reflect a blood and tissue copper overflow into the blood due to hepatic damage. This test has, however, only been evaluated in small groups and further validation will be required, particularly its specificity in other causes of chronic liver disease. The convenience of a reliable serum marker for diagnostic purposes is highly desirable for use in WD work up.

A key clinical feature in the diagnosis of WD is the presence of KF rings, occurring in 100% of individuals with neurological disease and less frequently in those with liver disease (Taly 2007). The phenomenon arises as a consequence of copper deposition in the Descemets membrane and indicates that free copper has been released into the individual's circulation. Visualisation requires the use of slit-lamp amplification (Walshe 2011).

Due to the invasive nature of liver biopsy, this is no longer commonly undertaken for routine diagnosis of WD; previous studies have shown that up to 40% of individuals at presentation may have cirrhosis (Merle 2007). Early histological changes of WD are non-specific and represent a spectrum that may include hepatic steatosis, chronic hepatitis and fibrosis and may add to diagnostic delay.

Acute liver failure due to WD is an important diagnosis to make early, affecting both the management of the patient and also enabling screening and diagnosis of other family members (Ostapowicz 2002). Acute hepatic failure in WD gives rise to many characteristic biochemical and haematological abnormalities due to the toxic effect of an acute copper release from hepatocyte lysis. The laboratory findings of fulminant WD previously described have included Coomb's negative haemolytic anaemia, low serum alkaline phosphatase and increased aspartate to alanine aminotransferase ratios (Berman 1999; Korman 2008; Lee 1998; Wilson 1987).

Consensus guidelines for the diagnosis of WD exist; however, many of the criteria have not been formally evaluated and issues such as sensitivity and specificity for index tests remain to be fully explored.

The objective of this review is to determine the diagnostic accuracy of the index tests for WD. The index tests covered by this Cochrane review are caeruloplasmin, 24-hour urinary copper and hepatic copper content.

We will investigate whether index tests should be performed in all individuals who have been recommended for testing for WD, or whether these tests should be limited to subgroups of individuals. Firstly, we will test differences in cut-off values and assay types as these are likely to have the most influence on heterogeneity. Depending on the number of studies included, we will perform further subgroup analyses considering age, gender, ethnicity, study design, assay type, clinical presentation (hepatic or neurological presentation or both).

Types of studies

We will include prospective or retrospective, cohort or cross-sectional studies that assess the diagnostic accuracy of an index test in the clinical context of the diagnosis of WD. These will include those with WD or suspected WD versus a normal population but may also include heterozygotes. We will exclude studies that evaluate the index test in a normal population without a WD comparator group or in use of diseases other than WD.

Participants

Studies including children and adults of all ages with WD confirmed by the Leipzig criteria (Ferenci 2003). Studies that do not meet these criteria will be excluded.

Index tests

The diagnostic accuracy of caeruloplasmin, urinary copper and liver copper content will be evaluated for the diagnosis of WD.

Target conditions

Wilson's disease as defined by the Leipzig criteria (Ferenci 2003).

Reference standards

The clinical reference standard is the diagnosis of WD as outlined by the Leipzig criteria (Ferenci 2003).

Electronic searches

We will apply no language or publication date restrictions in our search methods.

We will identify relevant studies from the Cochrane Cystic Fibrosis and Genetic Disorders Group's Inborn Errors of Metabolism Trials Register using the term: Wilson’s disease

The Inborn Errors of Metabolism Trials Register is compiled from electronic searches of the Cochrane Central Register of Controlled Trials (CENTRAL) (updated with each new issue of The Cochrane Library), weekly searches of MEDLINE and the prospective handsearching of one journal - Journal of Inherited Metabolic Disease. Unpublished work is identified by searching through the abstract books of the Society for the Study of Inborn Errors of Metabolism conference and the SHS (Scientific Hospital Supplies) Inborn Error Review Series.

We will also search the following databases:

PubMed www.ncbi.nlm.nih.gov/pubmed (1966 to present) (Appendix 1);
CINAHL (EBSCO 1982 to present);
Embase (OVID 1982 to present);
The Cochrane Library www.cochranelibrary.com (The Cochrane Database of Systematic Reviews (CDSR) and CENTRAL database);
Science Citation Index via the Web of Science (1898 to current);
Web of Science’s Conference Proceedings Citation Index (CPCI);
British Library’s ZETOC http://zetoc.jisc.ac.uk/wzgw?db=etoc (1993 to date) for conference abstracts;
PROSPERO (International Prospective Register of Systematic Reviews) www.crd.york.ac.uk/prospero/search.asp.

Our search strategy for PubMed will be adapted for use in other databases.

We will also search Clinical Trials.gov (www.clinicaltrials.gov/ ) and the WHO International Clinical Trials Registry Platform (www.who.int/ictrp/en/) in an attempt to identify ongoing trials.

Searching other resources

The reference lists of all included articles and relevant systematic reviews will be reviewed to identify additional studies not found through the electronic review.

Selection of studies

Two authors will independently review the titles and abstracts of articles found in the Electronic searches for potentially eligible studies for review, if possible the full paper. The same two authors will independently assess the full manuscripts against the inclusion criteria and resolve any disagreements over eligibility through discussion. Where necessary, a third review author will be consulted to make a final judgement on inclusion of studies.

If multiple articles report the same study or study participants, we will include the primary reference as the article with the most relevant information or the largest number of participants (or both) and we will list all other secondary references under the primary reference of the study.

Data extraction and management

Two authors will independently extract the following data from published articles; discrepancies will be resolved by discussion or by consulting a third author.

• First study author and year (of primary reference)
  
• Study design and setting
  
• Participant recruitment or sampling methods
  
• Number of participants
  
• Clinical and demographic characteristics (age, clinical presentation (hepatic versus neurological), ethnicity)
  
• Details of index text (assay type, control, cut-off values)
  
• Details of the reference standard
  
• Methodological quality of included studies (seeAssessment of methodological quality)
  

We will create 2 x 2 tables for each index test described in each study, cross-tabulating index test results with presence of the target condition (reference standard) in the following format:

Where a study does not present all relevant data for creating a 2 x 2 table, we will contact the study authors directly to request this information. If information cannot be provided, we will retain the study in the narrative section of the review, but will not include it in meta-analysis.

Assessment of methodological quality

Two authors (AR, SN) will independently assess the methodological quality of each included study using the QUADAS-2 tool (Whiting 2011) as recommended by Cochrane. We will consult a third author if we can not resolve disagreements by discussion. The tool is made up of four domains:

• participant selection;
  
• index test;
  
• reference standard;
  
• participant flow and timing.
  

We will assess each domain in terms of risk of bias, with the first three domains also considered in terms of applicability concerns. We have detailed the components and signalling questions associated with each of the domains in an appendix (Appendix 2).

Statistical analysis and data synthesis

We will consider each index test and method separately in the analysis. For all included studies, we will use the data in the 2 x 2 tables (see Data extraction and management). We will calculate two statistics for each index test in each study.

Sensitivity = number of true positives / number of participants with the target condition present = TP/(TP + FN)

Specificity = number of true negatives / number of participants without the target condition present = TN/(TN + FP)

We will conduct exploratory analyses by plotting study specific sensitivity and specificity estimates on both a forest plot and in receiver operating characteristic (ROC) space.

Given the lack of validated cut-offs of the index tests (see Index tests), we expect variability in cut-off points chosen in the included studies. Therefore, we will meta-analyse pairs of sensitivity and specificity using the hierarchical summary ROC (HSROC) model (Rutter 2001), which allows for the possibility of variation in threshold between studies, while also accounting for variation within and between studies and any potential correlation between sensitivity and specificity. Where adequate studies are available to estimate all parameters of the HSROC model, we will assume a symmetrical shape to the summary ROC curve. As summary estimates of sensitivity and specificity are difficult to clinically interpret when thresholds of included studies are variable in the HSROC model, we will derive estimates of summary sensitivity, positive and negative likelihood ratios at fixed values (median, lower and upper quartile) of specificity from the HSROC models. For these analyses we will use the NLMIXED procedure in SAS software (SAS 2011, version 9.3; SAS Institute 2011, Cary, NC).

Investigations of heterogeneity

We plan to investigate the following subgroups:

• age (to include all ages);
  
• gender;
  
• ethnicity;
  
• clinical features (hepatic and or neurological);
  
• index test method;
  
• different study designs.
  

In exploratory analyses, we will visually examine forest plots of sensitivity and specificity for each index test, and summary ROC plots to explore the effect of each of the factors of interest. If there are sufficient studies, we plan to perform meta-regression by including each potential source of heterogeneity as a covariate in the HSROC model.

Sensitivity analyses

If appropriate, we will perform sensitivity analyses excluding studies which are at a high risk of bias for at least one domain of the QUADAS-2 tool (see Assessment of methodological quality).

Assessment of reporting bias

We will not formally investigate reporting bias via existing analytical tools such as funnel plots due to current uncertainty around interpretation of such tools in this setting. Instead, we will perform systematic electronic searches and detailed searches of other published and unpublished sources (see Electronic searches and Searching other resources) in order to retrieve as many eligible studies as possible for inclusion in the review.

If an eligible study does not provide sufficient published information for us to include it in a meta-analysis, we will make every attempt to contact study authors to retrieve relevant data to make this possible.

Summary of findings of the review

We plan to summarise the results of the review in a summary of findings table, with results for all index tests summarised in the same table if appropriate. We will summarise the following information in this table: participants or population; prior testing; settings; index test and method; reference standard; target condition; importance; included studies; methodological quality; limitations and analysis results by index test. We will present analysis results in a manner which is appropriate for the analysis we are able to perform.

("Hepatolenticular Degeneration"[Mesh] OR hepatolenticular degeneration OR hepatic lenticular degeneration OR “wilson disease” OR “wilsons disease” OR “wilson’s disease”) AND (caeruloplasmin OR apocaeruloplasmin OR apoceruloplasmin OR holocaeruloplasmin OR holoceruloplasmin OR holo-caeruloplasmin OR holo-ceruloplasmin OR “urinary copper” OR “urine copper” OR “hepatic copper” OR “liver copper”).

QUADAS-2 is structured so that four key domains are each rated in terms of the risk of bias and the concern regarding applicability to the research question (as defined above). Each key domain has a set of signalling questions to help reach the judgments regarding bias and applicability.

Domain 1: Participant selection

Risk of bias - could the selection of participants have introduced bias?

1. Was a consecutive or random sample of patients enrolled?

2. Was a case-control design avoided?

3. Did the study avoid inappropriate exclusions?

Inappropriate exclusions include participants with tremor and chronic liver disease of unknown cause etc.

4. Could the selection of patients have introduced bias?

5. Concerns regarding applicability

Is there concern that the included participants do not match the review question?

Domain 2: Index test(s)

Describe the index test and how it was conducted and interpreted - this will vary with each test method. If more than one index test was used, please complete for each test.

Risk of bias - could the conduct or interpretation of the test have introduced bias?

1. Were the index test results interpreted without knowledge of the results of the reference standard?

2. If a threshold was used, was it pre-specified?

3 Could the conduct or interpretation of the index test have introduced bias?

4. Concerns regarding applicability

Are there concerns that the index test, its conduct, or interpretation differ from the review question?

Domain 3: Reference standard

Describe the reference standard and how it was conducted and interpreted.

Risk of bias - could the reference standard, its conduct, or its interpretation have introduced bias?

1. Is the reference standard likely to correctly classify the target condition?

2. Were the reference standard results interpreted without knowledge of the results of the index test?

Does not apply as the Leipzig criteria is a clinical reference standard and its application therefore will depend on application of results of index tests.

3. Could the reference standard, its conduct, or its interpretation have introduced bias?

Not sure this applies as the Leipzig criteria is a clinical reference standard and its application therefore will depend on application of results of index tests.

4. Concerns regarding applicability

Are there concerns that the target condition as defined by the reference standard does not match the question?

Domain 4: Flow and timing

· Describe any participants who did not receive the index test(s) and/or reference standard or who were excluded from the 2 x 2 table (refer to flow diagram).

· Describe the time interval and any interventions between index test(s) and reference standard - again, not sure this applies as it is a clinical reference standard.

Risk of bias - could the participant flow have introduced bias?

1. Was there an appropriate interval between index test(s) and reference standard?

Does not apply - clinical reference standard.

2. Did all participants receive a reference standard?

3. Did participants receive the same reference standard?

4. Were all participants included in the analysis?

5. Could the participant flow have introduced bias?

AR and SN drafted this protocol with input from PC. The final version of the protocol was agreed by all authors.

Aidan Ryan: none known.
Sarah Nolan: none known.
Paul Cook: none known.

• No sources of support supplied
  

• National Institute for Health Research, UK.
  
  This systematic review was supported by the National Institute for Health Research, via Cochrane Infrastructure funding to the Cochrane Cystic Fibrosis and Genetic Disorders Group.