Diagnosis and Outcomes of Late‐Onset Wilson's Disease: A National Registry‐Based Study

Wilson's disease (WD) is usually diagnosed in children and young adults; limited data exist on late‐onset forms.

Wilson's disease (WD) is a rare autosomal recessive disorder of copper metabolism, with progressive copper overload in many organs, primarily the brain and liver. 1 The diagnosis of WD involves a combination of clinical, biochemical, and genetic tests. 2 WD typically presents in children, adolescents, and young adults. The mean age at diagnosis ranged from 18.5 AE 11 3 to 28 AE 9 years 4 in different registry-based studies, with a later diagnosis in neurological forms. 5 Late-onset WD forms diagnosed after age 40 years, or even after 70 years, are rarely reported; therefore, little is known about this phenotype. 6,7 To date, a few case reports [7][8][9][10][11][12][13][14] and a European study (n = 46), with limited data on neurological phenotyping, have described patients with late-onset WD. 15 In this current study, we analyzed a cohort of patients diagnosed after age 40 years, who were identified from the French Wilson's Disease Registry (FWDR). Our aim was to characterize their clinical and paraclinical presentations, as well as their therapeutic management and outcomes.

Patients and Methods
This study was carried out by the National Reference Centre for Wilson's Disease (NRCWD) in Paris, working in collaboration with the French WD network. All patients had given their informed consent for genetic analysis and anonymous study of their data before their inclusion in the FWDR. The present study was approved by the Institutional Review Board (IRB00003888, IORG0003254, FWA00005831) for the French Institute for Medical Research and Health (INSERM) (no.: 19-550).

Study Design
Patients diagnosed with WD after age 40 years between June 1974 and 2016 were selected from the FWDR for this longitudinal cohort study. Patients had confirmed WD with a Leipzig score ≥4. 16 They were classified into three groups according to their diagnostic mode: (1) neurological (hepatic features could be associated); (2) hepatic (elevated liver enzymes, hepatomegaly/steatosis at liver ultrasound, compensated cirrhosis, and decompensated cirrhosis with or without encephalopathy); and (3) diagnosed by family screening or incidentally.

At Diagnosis
Sex; past medical history; age at first symptoms and at diagnosis; initial phenotype; clinical examination; and laboratory, ophthalmological, genetic, and imaging findings were reported. Neurological signs (tremor, writing difficulties, gait disorder, dysarthria, drooling, dystonia, cerebellar syndrome, parkinsonism, and cognitive impairment) and psychiatric disorders (depression, anxiety, psychosis, or behavioral disorders) were recorded. Slit lamp examination searched for the presence of Kayser-Fleischer ring (KFR). 17 Hepatic assessment included laboratory tests (blood count, liver enzymes, and prothrombin time), presence of hepatic copper if available, liver ultrasonography, and/or transientelastography-Fibroscan to determine the hepatic score from 0 to 6 (progressive grading ranging from no liver abnormalities to decompensated cirrhosis). 18 When liver transplantation was performed, the hepatic score was rated as 0 during follow-up. Copper tests included serum total copper, serum ceruloplasmin (immunonephelometric method), and 24-h urinary copper excretion (UCE). Non-ceruloplasmin-bound copper (NCC) was calculated as follows: NCC (μmol/L) = serum copper (μmol/L) À 0.047 Â ceruloplasmin (mg/L). 19,20 For patients diagnosed after 2010, the direct assay of NCC (exchangeable copper) and calculation of the relative exchangeable copper (REC) 18,21-23 were recorded. The REC is considered a valuable biomarker for WD diagnosis, whatever its clinical form, 18,[21][22][23] and is defined as the ratio of the direct assay of exchangeable copper to total serum copper.
Diagnosis was confirmed using molecular analysis of the ATP7B gene.
Brain images were obtained using a 1.5-T brain magnetic resonance imaging (MRI) with fluid-attenuated inversion recovery (FLAIR), T2, T1, and T2* sequences. All brain MRIs were reviewed by two experts from the NRCWD (A.P. and C.N.), who assessed T2/FLAIR hyperintensities. 18 In neurological patients, MRIs were also scored using the method of Dusek et al. 24 All treatments (D-penicillamine, trientine 2HCl, zinc salts, and liver transplantation) and their tolerance were reviewed.

At the Most Recent Follow-Up
Duration of follow-up, age, phenotype, clinical features, presence of KFR, brain MRI score, hepatic score, serum and urinary copper levels, and treatment were documented. Incidence and causes of death were recorded.

Statistical Analysis
The data are presented as median values with interquartile ranges (IQR) for continuous variables and numbers (percentages) for qualitative variables. All analyses were performed using statistical programming language R, version 4.0.3 (R Project for Statistical Computing). A nonparametric Wilcoxon test was performed to compare two distributions, whereas a nonparametric Kruskal-Wallis test was performed to compare three distributions. A χ 2 test or Fisher's test was used to test the relationships between categorical variables, based on the conditions of test validity. Box plots were used to represent copper findings at diagnosis and at follow-up by clinical presentation. All tests were bilateral and performed using a 2-tailed level of significance set at P < 0.05.

Results
A total of 552 patients with WD were included in the FWDR. Forty-five (8%) were at least aged 40 years at diagnosis (24 were aged 40-50 years, and 21 were ≥50 years). There were 20 men (44%). The median age at symptom onset was 45 years (IQR: 42-50) and at diagnosis 49 years (IQR: 43-53). The oldest patient was aged 64 years at diagnosis.
Twenty patients (44%) had a neurological presentation, 13 (29%) had a hepatic presentation, and 12 (27%) were diagnosed by family screening or incidentally (Table 1). Age at first symptoms, age at diagnosis, and time to diagnosis were not significantly different between the neurological and hepatic groups. Some previous manifestations were suggestive of WD: depression (n = 4), bipolar disorder (n = 1), anorexia nervosa (n = 2), unexplained transient hepatitis or chronic liver disease (n = 3), and cytopenia (n = 1). Figure 1 shows copper findings between baseline and the most recent follow-up. At baseline, the 24-h UCE values were normal in 7 of 21 patients (33%), including 4 from the screened group. The 24-h UCE values differed significantly between the groups (P = 0.04), with higher levels among hepatic patients.
The sensitivity of REC for WD diagnosis was 82% in symptomatic patients and 86% in screened patients. The test failed to diagnose WD in 3 patients with hepatic form (all had cirrhosis, high UCE, and low ceruloplasmin), in 1 patient with atypical neurological presentation and in 2 screened patients. The REC was greater than 15% in 5 of 7 patients with a normal UCE and between 8% and 15% for the remaining 2 of 7 patients.
Full genetic data were available for 39 of 45 patients (87%). Two mutations of the ATP7B gene were found in all patients except 1 patient (one mutation found) who had a classic neurological form. Most patients carried heterozygous mutations (28 of 39); the most frequent mutation was H1069Q.
The initial treatment and tolerance are summarized in Table 2. Most patients received D-penicillamine as firstline therapy (57%). There was no significant difference in treatment tolerance. Eighteen patients (41%) needed a second-line therapy, mainly due to lack of efficacy. Table 3 presents patients' characteristics at the most recent follow-up, 6.5 years (median) after diagnosis. Seven patients were lost to follow-up, and 2 died.

Neurological Group (n = 20)
At diagnosis (Tables 1 and 4). Fourteen neurological patients (70%) presented typical signs of WD, either isolated or associated, and had a median diagnostic delay of 19 months (IQR: 11.5-33). In all 10 patients whose brain MRIs were available, imaging was abnormal with FLAIR hyperintensities in the basal ganglia, thalamus, mesencephalon, and pons. KFR was present in all these classic neurological patients. The median hepatic score was 1/6, including only 1 patient with cirrhosis. Three patients had thrombopenia.
Six patients from the neurological group (30%) presented with isolated long-lasting symptoms of WD (Table 4), with a median delay of 120 months before diagnosis (IQR: 12-120). One patient, aged 47 years, presented with isolated torticollis for 10 years, treated at some point with botulinum toxin. He subsequently developed writer's cramp, dysarthria, cerebellar ataxia, and mild tremor of the right hand, which enabled the diagnosis of WD. Two brothers were in this group: one had isolated writer's cramp for a year before being diagnosed with WD at age 61 years, when he was also presenting with cirrhosis; the other one (aged 58 years) had writer's cramp with hand tremor for 10 years before being diagnosed. Two patients presented with functional movement disorders (FMD), based on Gupta and Lang diagnostic criteria (eg, inconsistency in movements over time, incongruence with typical features of movement disorders, and suggestibility) 25 : one had paroxysmal myoclonus of the limbs and the head, with normal neurological examination between the episodes of abnormal movement, whereas the other had an inconsistent and variable tremor of both hands and head, associated with balance difficulties. Electrophysiological analysis of both patients provided further support for FMD diagnosis. The first one displayed muscular jerks of long duration whose diffusion did not follow either the cortico-spinal or the proprio-spinal descending pathway, with variability in the contraction interval of the different muscles and of their activation sequence. The analysis of the second patient found no tremor or other abnormal movement. The sixth patient had isolated dysarthria for almost 10 years, with progressive worsening over time, and then developed tremor of upper limbs and walking problems. Surprisingly brain MRI of all 6 patients showed no T2/FLAIR hyperintensities (Fig. S1); however, the brain MRI severity scale, according to the recent validated score by Dusek et al, revealed mild atrophy in 4 of them 24 (Table S1). Only 1 of these 6 patients had a KFR. Psychiatric symptoms included depression (n = 7) and behavioral disorder (n = 2, neurological patients). c Three patients with decompensated cirrhosis had available MELD scores (32, 18 and 29).  high intrahepatic copper. All these atypical neurological forms of late-onset WD were confirmed by molecular analysis (Table S2). Three mutations were novel: their pathogenicity was assessed using the American College of Medical Genetics and Genomics classification, 26 based on general population databases, missense variant prediction algorithms/software, international databases of pathogenic variants, and the literature.
After 5.0 Years (IQR: 3.0-12.8) of Follow-Up. Follow-up was available for 19 patients with neurological presentation ( Table 3). All patients were treated, including 12 of 19 with D-penicillamine (  19). The REC values decreased significantly in the neurological patients at the most recent follow-up (P = 0.002), unlike hepatic and screened patients (Fig. 1).
Five of the 6 patients with atypical neurological presentations were clinically stable or improved (Table 4). Of these 6 patients, brain MRI remained normal in 3 of them, worsened slightly in 1, and was not repeated in 2.
Overall, 3 patients with neurological forms worsened despite treatment. Two patients from the classic presentation group were treated using chelators and had a progressive worsening of their parkinsonism and dysarthria after initiation of chelation. The third one belonged to the atypical group and developed copper deficiency 20 years after initiation of treatment, resulting in severe chronic sensorimotor polyneuropathy of the lower limbs (Table 4, patient 3). At that time, his hepatic score was stable (1/6), UCE and exchangeable copper levels were low, and brain and medullar MRI findings remained normal. Other causes of polyneuropathy were ruled out. After zinc acetate was withdrawn, the patient slowly improved, although after 8 years he still experienced deep and superficial sensory disturbances, writer's cramp, and dysarthria. He is no longer being treated for WD but undergoes regular clinical and laboratory tests. His liver status has remained stable.

(Continues)
MRI available had hyperintensities in the basal ganglia, suggestive of neuropathological changes secondary to copper deposition. Two patients with encephalopathy displayed severe underlying neurological signs of WD, 1 of whom had a 1-year history of psychomotor impairment before the onset of hepatic decompensation. Examination by a neurologist 3 months later found axial dystonia, risus sardonicus, blepharospasm, and dysexecutive frontal syndrome. Neurological examination of another patient revealed signs of central pontine myelinolysis and postintensive care neuropathy. Brain MRI findings in these 2 patients were abnormal. After 4 Years (IQR: 3.5-9) of Follow-Up. Hepatic score improved in 70% of patients at last follow-up (Table 3). Ultrasound abnormalities improved in all patients. One patient who was initially encephalopathic displayed persistent dysarthria, swallowing disorder, parkinsonism, walking problems, frontal syndrome, and mood disorder. KFR disappeared in 3 of 6, and brain MRI improved in 2 of 3 patients with available follow-up data.
Between baseline and the most recent follow-up, serum ceruloplasmin levels increased significantly (P = 0.031) and exchangeable copper decreased significantly (P = 0.011), in contrast to neurological patients and screened patients (Fig. 1).
Five patients required a second-line therapy after a median of 9 months (Table 2). Four patients underwent liver transplantation as first-line therapy and 2 patients as second-line therapy due to deterioration (cirrhotic decompensation and acute liver failure).

Screened Patients (n = 12)
At diagnosis (Table 1). Eleven patients were diagnosed by family screening, and 1 was diagnosed fortuitously when being assessed for recurrent vasovagal malaise, for which copper tests were performed. She had a normal brain MRI, low serum copper and ceruloplasmin, and normal UCE at diagnosis. Three of 7 patients for whom full data were available were related to other patients with late-onset WD. When available (n = 3), brain MRI was normal. KFR was present in 2 of 10 patients. Liver assessments were abnormal in 3 of 7 patients, with minimal liver ultrasound abnormalities. Three patients with normal copper balance were not treated but followed a low-copper diet.

Discussion
In this national registry-based study, 8% of patients with WD had a late-onset form diagnosed after age 40 years. This is twice as much as that reported in the 2007 European study. 15 This discrepancy may be due to improved family screening and the discovery of new mutations. The median age at diagnosis was 49 years, in line with the data of the European study. 15 Most of the patients presented with neurological symptoms. Those with a hepatic presentation displayed severe forms: a great majority had cirrhosis, and 46% of them required liver transplantation.
The most intriguing result of our study was that 30% of the patients with late-onset neurological WD presented with atypical neurological symptoms such as long-standing isolated dysarthria, focal dystonia (cervical dystonia or writer's cramp), or FMD (functional tremor/myoclonus confirmed by electrophysiological analysis). In the literature, one case report describes a 25-year-old man with WD presenting with paroxysmal dystonia over several years. 27 The association between functional symptoms and neurological disorders is increasingly reported. A recent large multicenter study of 410 patients with functional motor disorders showed that 22% had comorbid neurological diseases. 28 In these patients, functional symptoms could appear before the diagnosis of the neurological condition. 28 In Parkinson's disease, associated functional neurological disorders are not rare 29 and are typically expressed on the side most affected by parkinsonism. 30 The association between functional and "organic" disorders may be due to common neurophysiological abnormalities. [30][31][32] In late-onset WD, these presentations are very misleading, especially because none of the 6 patients with atypical neurological form had brain MRI hyperintensities suggestive of WD (however, atrophy was present in 4 of them). Moreover, most of them had no KFR. In contrast, in classic early-onset WD cohorts, more than 90% patients with neurological forms have abnormal brain MRI. 24,33 The UCE and REC were helpful to guide the diagnosis toward WD before genetic confirmation. These peculiar lateonset neurological presentations have not previously been described to our knowledge. They appear different from purely psychiatric presentations of WD in which brain imaging findings can be normal. It is difficult to know if neurological symptoms were due to copper toxicity in these patients. Copper was not available in the cerebrospinal fluid. However, it has been shown in an anatomopathological study that copper can be increased in the brain even in pure hepatic forms without neurological symptoms or brain imaging abnormalities. 34 Thus far, neither the existence of modifier genes nor specific mutations of the ATP7B gene are known to be linked to such atypical neurological phenotypes. 35 Interestingly, one of the responsible mutations for WD in an atypical case (homozygous Met645Arg variant) has recently been shown to cause exon skipping, resulting in 30% wild-type expression, 36 and may explain the late presentation. One may wonder whether some of the mutations lead to slower copper deposition in the brain and cornea. Longitudinal follow-up of these patients will help us to specify these peculiar presentations. These data should encourage neurologists to consider a diagnosis of WD more often in patients after age 40 years. A history of psychiatric disorders (depression, bipolar disorder, and anorexia nervosa) and laboratory abnormalities (isolated thrombocytopenia and elevated liver enzymes) were suggestive of WD in a few patients and should be considered as red flags. 37 The diagnosis of late-onset neurological WD seems therefore challenging. In our study, the median diagnostic delay of neurological patients was longer (20 months; range: 0-180) than that in the classic early-onset forms 3-5 (13.9 months among patients in the FWDR 3 ). In the European late-onset cohort, 15 patients with neuropsychiatric WD had a median diagnostic delay of 2.8 years (range: 0-15). The specific organization of the WD network in France, with dedicated WD centers throughout the country, may explain these differences between the European and French cohorts.
Seventy percent of neurological patients developed typical signs of neurological WD (dystonia, tremor, dysarthria, and drooling) associated with abnormal brain MRI findings and KFR and liver disease in 75%. Cirrhosis was infrequent (13%) but may have been underestimated as liver biopsy was not performed. This data should encourage neurologists to look systematically for liver involvement (liver biology and liver ultrasound at least) in neurological WD patients. The diagnostic features and the evolution of neurological patients were similar to those described in the classic early-onset forms. 5 In hepatic patients, a high rate of cirrhosis (77%) was reported at diagnosis, including 3 cases of fulminant decompensation with encephalopathy. The severity of liver damage at diagnosis should encourage physicians to screen widely for WD in the event of liver disease of unknown etiology, regardless of age, even in the case of severe presentation (eg, decompensated cirrhosis and acute liver failure). 10 This is worthwhile because treatment is highly effective and hepatocellular carcinomas are not rare. In the hepatic group, the median value of REC was 27.6%, and KFR was present in more than 50% of patients, which did not differ from classic forms. 38 Our findings were similar to that of Ferenci et al in late-onset forms. 15 Brain MRIs were not always routinely performed in hepatic patients. Some of them had minimal neurological signs at diagnosis, which were not identified until later. Similarly, screened patients did not have a systematical screening for KFR or brain abnormalities. This highlights the need for multidisciplinary management of WD, particularly at diagnosis and in atypical forms.
When WD is suspected, laboratory tests are systematic. The previous European study in patients with lateonset WD 15 reported that the diagnostic criteria and genetic tests were the same as those for classic forms. In our study, UCE levels were significantly higher in hepatic patients than in the other groups, which has been previously described. 39 However, the UCE value could be normal, unlike ceruloplasmin and serum copper levels which were always low, increasing the difficulty in diagnosing atypical neurological forms. The REC is already known to have high sensitivity and specificity in diagnosing classic forms. [21][22][23] In this late-onset WD study, REC was less powerful than in the classic forms; results should be confirmed in a larger cohort.
As laboratory findings may be misleading, a confirmation of late-onset WD could come only from molecular analysis of the ATP7B gene. Two mutations were reported in all patients with available data, except 1 who had only one mutation. False negatives are possible if Sanger sequencing is used alone without multiplex ligationdependent probe amplification or next-generation sequencing. 40,41 The most frequently detected mutation was H1069Q, as described in all forms among Caucasians 41 and in the late-onset European cohort. 15 In France, all first-degree relatives of a patient with WD are screened for the disease, including family members above age 40 years. This allows the identification of about 20% of patients (French National Registry data).
Ten patients worsened in our study, mainly neurological patients and those on D-penicillamine, which is similar to what has been described in classic WD cohorts. 42 Zinc salt and trientine 2HCl were effective on neurological symptoms as first-line therapy with good tolerance, as in classic forms. 43,44 Liver disorders frequently worsened in screened patients (43%), in contrast to the other groups. Poor adherence to treatment in these asymptomatic patients may explain this deterioration. 45 In our cohort, 3 screened patients followed only a low-copper diet and did not worsen after 7 years of follow-up. Simple but regular follow-up visits could therefore be discussed in the exceptional cases of asymptomatic patients with normal copper levels, although an expert opinion is crucial in this decision. This is justified by the risk of treatment-induced copper deficiency, which can cause severe chronic sensorimotor polyneuropathy and myelopathy. 46,47 If the UCE and exchangeable copper values decrease below the normal threshold on therapy in association with leucopenia and anemia, physicians should consider lowering dosages or even discontinuing treatment with close monitoring.
Our study bears limitations. Its retrospective nature and the small number of patients limited our data analyses. However, the organization of the FWDR allowed us to obtain standardized and good-quality data. The rarity of WD itself does not allow for prospective studies with reasonably large sample sizes. In addition, three mutations in the patients with atypical neurological WD forms were novel; although in silico prediction tools suggested their pathogenic effect, gene expression studies are yet required to confirm their pathogenicity.
In conclusion, WD should be considered regardless of age, particularly because its presentation can be highly heterogeneous, with brain MRI abnormalities limited to atrophy and no KFR. Genetic testing confirms the diagnosis and should be performed in all first-degree members of a patient with WD. The prognosis of late-onset forms seems to be good when the treatment is personalized.