Assessment of Diagnostic Value of Serum Ceruloplasmin for Wilson’s Disease in Children

Serum ceruloplasmin is one of the major diagnostic parameters for Wilson’s disease (WD). Age and gender difference of serum ceruloplasmin remain controversy. This study aims to assess diagnostic value of serum ceruloplasmin for WD in children. Serum ceruloplasmin were measured in 317 WD patients, 21 heterozygotes, 372 healthy control children and 154 non-WD patients. Receiver operating characteristic (ROC) curve was used to determine the diagnostic accuracy of serum ceruloplasmin for WD in children. may provide the highest accuracy for diagnosis of WD in children.


Introduction
Wilson's disease (WD) is an autosomal recessive disorder of copper metabolism caused by mutations in the ATP7B gene [1]. Failure of copper transport from hepatocytes into bile results in copper accumulation in the liver and extra-hepatic organs. If WD is not recognized and treated early, hepatic and neurologic damage may be rapid and fatal. Therefore, early diagnosis is crucial to prevent the progression of the disease and to approach a good outcome of WD.
Ferenci et al proposed a conventional diagnostic scoring system for WD, including low serum ceruloplasmin levels, elevated urinary copper excretion, KF ring, liver copper content and genetic ndings [2]. Socha et al. further published a position paper of speci c diagnostic criteria for WD in children. With the application of Ferenci scoring system, WD can be identi ed and diagnosed easily when the patients present with typical symptoms. However, the diagnosis of WD is not easy in asymptomatic children, and in patients with acute liver failure [3][4][5].
Ceruloplasmin is the main copper-carrying protein in blood. With the defect of copper transportation, noncopper bound apo-ceruloplasmin is rapidly degraded [6]. Decreased level of serum ceruloplasmin is the most prominent biochemical nding in WD patients. According to the diagnostic consensus, serum ceruloplasmin < 20 mg/dL is conventionally considered as one of the major diagnostic threshold for WD [2]. Socha et al. suggested that WD should be highly suspected with serum ceruloplasmin < 10 mg/dl in children [7]. However, serum ceruloplasmin may vary with age, change with in ammation or estrogen, and decrease in some heterozygous carriers or patients with liver failure [8,9]. There have been few studies evaluating the diagnostic accuracy of serum ceruloplasmin in WD children [10][11][12][13]. Serum ceruloplasmin is one of the major diagnostic parameters for WD, but there have been few information about the diagnostic value of serum ceruloplasmin in WD children [14]. Therefore, it is important to establish age and gender speci c reference of serum ceruloplasmin in healthy children, and assess the diagnostic criteria of serum ceruloplasmin for WD during early childhood.
In our previous study, we have noticed that serum ceruloplasmin is decreased in the majority of WD children [15]. This study aims to provide reference value of serum ceruloplasmin in healthy children strati ed by age and gender, and assess diagnostic criteria of serum ceruloplasmin for screening and early diagnosis of WD in children.

Study subjects
A total of 317 children with WD diagnosed at Guangzhou Women and Children's Medical Center from January 1, 2010 to December 31, 2020 were enrolled in this study (n = 317). The diagnosis of WD was established on a total of score of four or more according to the consensus for WD [2,7]. WD patients were categorized into three groups based on clinical manifestations at diagnosis: (1) Asymptomatic WD (n = 261);(2) WD with acute liver failure (n = 15); (3) other types of WD, including acute hepatitis, chronic hepatic disease, neurological disease, renal disease and purpura (n = 41).
Heterozygous group (n = 21) included unaffected siblings carrying one pathogenic mutation in the ATP7B gene. Non-WD patients were represented by pediatric patients of GWCMC with exclusion of WD, including acute liver failure by infection or acetaminophen (n = 20), viral hepatitis (n = 38) and nephrotic syndrome (n = 96). Data of the patients with non-WD diseases were retrieved from database of clinical lab in GWCMC.
Healthy controls (n = 372) included 360 children who came for routine physical examination and 12 wildtype siblings of WD patients. Blood samples in healthy controls were originally collected for liver function test and then used for further measurement of serum ceruloplasmin. This study was approved by the ethical clinical research committee of Guangzhou Women and Children's Medical Center, and appropriate informed consents were obtained parental permission for all participants.

Biochemical And Genetic Analysis
The clinical variables of WD patients and their siblings, including age sex, family history, clinical manifestations and physical examinations, were collected from participant medical history and chart record at diagnosis. Serum ceruloplasmin were measured by immunonephelometry using the Beckman Coulter Immage 800 with standard reagents (Beckman Coulter, Brea, CA, USA).
Genetic analyses for the ATP7B gene mutation were performed in all WD patients and their siblings. Genomic DNA was extracted from peripheral blood samples. DNA sequencing was applied to the entire coding regions of exons 1-21 of the ATP7B gene with adjoining intron boundaries. The exported sequences were aligned and inspected with the reference sequence using DNAMAN software.

Statistical analysis
Statistical analyses were performed using SPSS (version 25; SPSS Inc., Chicago, IL, USA) and GraphPad Prism (version 7.0; GraphPad Software, La Jolla, California, USA). Continuous quantitative variables (age, serum Ceruloplasmin) were expressed as the means and standard deviations (SD). Sensitivity, speci city, positive predictive value (PPV) and negative predictive value (NPV) were calculated for assessment of diagnostic accuracy. The receiver operating characteristic (ROC) curve and the corresponding area under the curve (AUC) were used to determine the optimal cutoff value of serum ceruloplasmin for diagnosis of WD. A value of P < 0.05 was considered statistically signi cant.

Serum ceruloplasmin in healthy children
Serum ceruloplasmin were measured in 372 healthy children from the age of newborn to 15 years. Of these participants, 218 (58.6%) were boys, with a mean age of 6.8 ± 4.2 years and 154 (41.4%) were girls, with a mean age of 6.2 ± 3.8 years. The mean level of serum ceruloplasmin in these healthy controls was 30.7 ± 7.8 mg/dL. Among them, 30 (8.1%) had serum ceruloplamin < 20 mg/dL, whereas 13 (3.5%) had serum ceruloplasmin < 15 mg/dL. Age speci c reference value of serum ceruloplasmin was analyzed (Table 1 & Fig. 1). These healthy children were partitioned into 16 age subgroups. Serum ceruloplasmin was at the lowest level (25.5 ± 10.1 mg/dL) in children younger than 6 months, then rapidly increased to 33.0 ± 9.3 mg/dL by the age of 1 year, and maintained 29-34 mg/dL from 1 to 15 years old. There was no age speci c difference of serum ceruloplasmin in healthy children. Table 1 Serum ceruloplasmin in healthy children strati ed by age Gender speci c reference value of serum ceruloplasmin was also analyzed in these children (Table 1 & Fig. 1). The mean levels of serum ceruloplasmin in healthy boys and girls were 31.8 ± 7.7 mg/dL and 29.2 ± 7.6 mg/dL respectively, exhibiting statistically signi cant gender difference (p < 0.01). While analyzing gender difference in each age subgroup, serum ceruloplasmin also exhibited higher level in boys than girls, but there was no signi cance with the exception of 8-year subgroup.
Serum Ceruloplasmin In Children With Wilson's Disease As shown in Table 2, a total of 317 children were diagnosed as WD in GWCMC in last 11 years, 200 (63.1%) boys and 117 (36.9%) girls. The ratio of male:female in WD patients was 1.7:1. At diagnosis, the mean age of these patients was 6.1 ± 2.8 years (ranging, 0.6-15.0 years). The mean level of serum ceruloplasmin of them was 5.7 ± 4.7 mg/dL, which was signi cantly lower than that in healthy control subjects (p < 0.001). However, 6 (1.9%) of 317 WD children had normal value of serum ceruloplasmin > 20mg/dL, and they were all asymptomatic (6/261, 2.3%).  Fig. 1). The youngest WD patient was a 7-month-old girl diagnosed through family screening. She was asymptomatic with low serum ceruloplasmin of 4 mg/dL but normal liver function. Among 317 WD patients, the mean level of serum ceruloplasmin varied in different age subgroups. No age speci c difference of serum ceruloplasmin was shown.
Gender speci c difference of serum ceruloplasmin in WD patients was also analyzed. Of the 317 WD patients, the mean age of boys was 6.2 ± 2.8 years, and that of girls was 5.9 ± 2.9 years. There was no signi cant gender difference of serum ceruloplasmin among 317 WD patients and inside each age subgroup (Fig. 1). However, among 261 (82.3%) asymptomatic WD patients, serum ceruloplasmin in 164 boys (6.0 ± 5.1 mg/dL) was signi cantly higher than that in 97 girls (4.6 ± 3.7 mg/dL) (p < 0.05), exhibiting similar trend with that in healthy children.
The association between clinical manifestation and the level of serum ceruloplasmin in WD patients was analyzed (Table 3). Of 261 asymptomatic WD children, serum ceruloplasmin was signi cantly lower and the age at diagnosis was much younger while compared with 56 WD children with various clinical symptoms and signs (p < 0.05, p < 0.001). Of 15 WD children with acute liver failure, serum ceruloplasmin was signi cantly higher and the age at diagnosis was much older while compared with 302 WD children without acute liver failure (p < 0.01, p < 0.001). None of WD patients with acute liver failure had serum ceruloplasmin > 20 mg/dL.

Serum Ceruloplasmin In Patients With Non-wd Diseases
Serum ceruloplasmin was measured in 154 patients with non-WD diseases, including acute liver failure, viral hepatitis and nephrotic syndrome. The mean age of these non-WD patients was 4.9 ± 3.7years (ranging from 1 month -15.3years) ( Table 4). All non-WD patients had serum ceruloplasmin > 20 mg/dL, with the mean level of serum ceruloplasmin (28.6 ± 11.8 mg/dL) much higher than that in WD patients (p < 0.001), but lower than healthy controls (p < 0.001). Serum ceruloplasmin in WD patients with R778L in the ATP7B gene Of the 317 WD patients, 313 had genetic testing for pathogenic mutations in the ATP7B gene, with 307 biallelic mutations and 6 one allelic mutation. Through family screening, 21 heterozygotes and 12 wild type siblings were identi ed. All wild type siblings were healthy and included in healthy controls. All heterozygotes were healthy without liver dysfunction. The mean level of serum ceruloplasmin in heterozygotes was (25.6 ± 5.9mg/dL), signi cantly lower than healthy controls (30.7 ± 7.8mg/dL, p < 0.05), but much higher than WD patients (5.7 ± 4.7mg/dL, p < 0.001) (Fig. 1).
Among 313 WD patients, the most frequent mutation in the ATP7B gene was R778L, identi ed in 33.9% (106 patients) of 313 WD patients (Table 5). Of the 313 WD patients, 9 carried bi-allelic R778L mutations (R778L homozygotes), 97 had one allelic mutation of R778L together with one allelic mutation other than R778L (R778L heterozygotes) and 207 had none of R778L (No R778L). The patients with R778L homozygotes had the lowest level of serum ceruloplasmin (2.3 ± 0.5mg/dL) and signi cantly lower than WD children without R778L (6.2 ± 4.8 mg/dL) (p < 0.05). Diagnostic Accuracy Of Serum Ceruloplasmin In Wilson's Disease The ROC curves were constructed by using the data of 317 WD patients, 21 heterozygotes, 372 healthy controls and 154 non-WD patients. As shown in Fig To de ne the accuracy of serum ceruloplasmin for the patients with acute liver failure, the ROC curve was constructed by using the data of 15 WD patients and 20 non-WD patients with acute liver failure. The area under the curve was 0.952 (95% con dence interval (CI), 0.885-1). The ROC curve indicated that the cutoff value of serum ceruloplasmin of 16 mg/dL gave the highest diagnostic accuracy for WD in the patients with acute liver failure (Fig. 2B).
To further de ne the accuracy of serum ceruloplasmin for diagnosis of WD in the patients with elevated aminotransferases, the ROC curve was constructed by using the data of 256 asymptomatic WD patients with elevated aminotransferases and 38 non-WD patients with viral hepatitis. The area under the curve was 0.994 (95% con dence interval (CI), 0.988-1). The ROC curve suggested that the cutoff value of serum ceruloplasmin of 16.85 mg/dL gave the highest diagnostic accuracy for WD with elevated aminotransferases (Fig. 2C).
As shown in Fig. 2D-F, the mean level of serum ceruloplasmin of 317 WD patients was 5.7 ± 4.7 mg/dL, signi cantly lower than that in all of 547 non-WD children (6.0 ± 4.0 mg/dL) (p < 0.001). For the patients with acute liver failure, the mean level of serum ceruloplasmin in WD patients (9.3 ± 4.7 mg/dL) was lower than that in non-WD patients (29.7 ± 14.3 mg/dL) (p < 0.001). In the situation with elevated aminotransferases, the mean level of serum ceruloplasmin in 249 asymptomatic WD patients (5.5 ± 4.7 mg/dL) was signi cantly lower than that in 38 patients with viral hepatitis (32.5 ± 7.9 mg/dL) (p < 0.001).

Discussion
To the best of our knowledge, the present study recruited the largest number of pediatric WD cases with 317 patients. The study also included 372 healthy children for study of age and gender speci c serum ceruloplasmin. As shown in Fig. 1, serum ceruloplasmin was at mildly low level only in healthy children < 6 months, but still much higher than that in WD children. These nding was consistent with previous reports, indicating that serum ceruloplasmin is not suitable for newborn screening [7,[16][17][18]. However, the present study indicated serum ceruloplasmin may be a good diagnostic criteria for WD children older than 6 months.
Gender difference of serum ceruloplasmin among healthy children was found in this study, which the boys had signi cantly higher level of serum ceruloplasmin than the girls. This trend in gender difference of serum ceruloplasmin was not only seen in all healthy children but also in different age subgroups (Fig. 1). Further, serum ceruloplasmin in asymptomatic WD patients exhibited the same trend of gender difference, which was higher in asymptomatic WD boys than girls (Fig. 1). There were very few information about gender difference of serum ceruloplasmin in children [9,19,20]. Previous study indicated that serum ceruloplasmin has no gender difference except being higher in girls during puberty due to estrogen stimulation [19]. Our study provided very different ndings in both healthy children and asymptomatic WD patients based on a large cohort of pediatric WD patients and healthy controls. Gender difference of serum ceruloplasmin should be considered at diagnosis of WD in childhood, and further investigation is needed.
It has been well known that diagnosis of WD is di cult in the asymptomatic children, and in the patients with acute liver failure [3][4][5]. Serum ceruloplasmin is one of the major diagnostic criteria, however, it has been reported that only 85% of the asymptomatic WD had diagnostic ceruloplasmin levels [21]. With 261 asymptomatic WD children (82.3% of 317 WD patients), this study assessed the diagnostic value of serum ceruloplasmin for WD and presented different nding. Serum ceruloplasmin was generally much lower in WD patients than healthy children, and only 6 asymptomatic WD children of all (6/261, 2.3%) had normal level of serum ceruloplasmin (> 20mg/dL). Our nding is consistent with the report from other studies from China and Korea, indicating that serum ceruloplasmin is a reliable criteria for early diagnosis of asymptomatic WD in children [6,12]. Some studies from Europe reported that about 30% of European WD patients had serum ceruloplasmin > 20 mg/dL [22][23][24]. The racial difference and genetic difference may be the reasons, but small size of the study may also be the other reason.
The ROC curve has been used to investigate the optimal cutoff value of serum ceruloplasmin for early diagnosis of WD. It is controversial for optimal cutoff value of serum ceruloplasmin for diagnosis of WD, which is from 11.5 mg/dL to 20 mg/dL [9][10][11][12][13]25]. It has been reported that in the patients with acute liver failure, the conventional cutoff of serum ceruloplasmin 20mg/dL only provided a diagnostic sensitivity of 21% and speci city of 84% for WD [22,25]. However, most of these studies had some degrees of bias in patient population, with either very few amount of pediatric WD patients or too much more WD patients with neurological symptoms, therefore may not provide an optimal cutoff value in children [3,12,25]. In the present study, the ROC curve analyses showed the optimal cutoff value of serum ceruloplasmin is 16.85 mg/dL, with a sensitivity of 95.9% and speci city of 93.6% for diagnosis of WD in children. Using the conventional cutoff value of serum ceruloplasmin 20mg/dL, the sensitivity is 98.1% and a speci city is 86.5%. Our nding was consistent with that in another Chinese study for WD adult patients, indicating that the threshold of serum ceruloplasmin of 16.8 mg/dL may be more accurate and favor higher specialty for diagnosis of WD in children as well as adults [12].
Genotypic difference of serum ceruloplasmin was also investigated in this study. The genotype of the ATP7B gene was associated with the level of serum ceruloplasmin ( Table 5). The patients with the genotype of R778L homozygotes had much lower serum ceruloplasmin level (2.3 ± 0.5mg/dL) than the patients without R778L (6.2 ± 4.8 mg/dL) (p < 0.05). Previous study reported that R778L homozygotes are associated with lower level of serum ceruloplasmin and the early onset of WD with hepatic presentation [26]. Other study reported that truncating mutations in the ATP7B gene are also associated with very low serum ceruloplasmin and an early onset of WD [27]. Taken together, genotype in the ATP7B gene is associated with onset of WD and the level of serum ceruloplasmin. Our study further revealed R778L homozygotes may cause decreased serum ceruloplasmin and hepatic damage at very early age.
In conclusion, serum ceruloplasmin is a reliable biochemical criteria for early diagnosis of WD in children older than 6 months. Gender and genetic difference of serum ceruloplasmin should be considered at diagnosis and treatment for WD patients. The cutoff value of serum ceruloplasmin 16.8 mg/dL may provide the highest accuracy for diagnosis of WD.

Declarations
Ethics approval and consent to participate The study was approved by the Medical Ethics Committee for Clinical Ethical Review, Guangzhou Women and Children's Medical Center ([2020] No.63901) and informed consent was obtained from a parent of each participant under 16 years old. The study was conducted according to Declaration of Helsinki.

Consent for publication
Not applicable.

Con ict of interests
The authors did not have con icts of interest to declare for this work.  Figure 1 Gender difference of serum ceruloplasmin was analyzed among healthy children, WD patients and heterozygotes. (A) There were no signi cant gender differences of serum ceruloplasmin intra each age subgroup among WD patients and among healthy children. (B) Serum ceruloplasmin was higher in boys than girls among healthy children (p<0.01) and also among WD patients (p<0.05). The mean levels of serum ceruloplasmin in WD patients were much lower than that in healthy controls and heterozygotes. Figure 2