Serum ceruloplasmin oxidase activity is a sensitive and highly specific ...

Journal of Hepatology 51 (2009) 925–930 www.elsevier.com/locate/jhep

Serum ceruloplasmin oxidase activity is a sensitive and highly specific diagnostic marker for Wilson’s diseaseq Uta Merle*, Christoph Eisenbach, Karl Heinz Weiss, Sabine Tuma, Wolfgang Stremmel Department of Internal Medicine IV, University Hospital, Im Neuenheimer Feld 410, 69120 Heidelberg, Germany

Background/Aims: A low serum ceruloplasmin concentration is considered diagnostic for Wilson’s disease. We aimed to evaluate an enzymatic test for ceruloplasmin oxidase activity and to compare it with the routinely used immunological ceruloplasmin measurement. Methods: Serum ceruloplasmin was measured enzymatically with o-dianisidine dihydrochloride as substrate and immunologically. 110 Wilson’s disease patients, 52 healthy controls, and 51 patients with impaired liver function not due to Wilson’s disease were analyzed. Assay performance was tested by receiver operating characteristic curve analysis, McNemar test, and Spearman’s rank correlation. Results: The greatest sum of sensitivity and specificity was seen for the enzymatic ceruloplasmin assay at a cut-off point of 55 U / L (93.6% and 100%, respectively) and for the immunologic assay at a cut-off point of 0.19 g /L (93.6% and 78.8%, respectively). For healthy controls, the differences in specificity between both assays were statistically significant (McNemar, p = 0.02). When additionally including patients with impaired liver function into the control group the specificity declined to 84.5% for the enzymatic assay and to 68.9% for the immunologic assay. The correlation between the enzymatic and immunologic assay was high in healthy controls (r = 0.94), but weaker in Wilson’s disease patients (r = 0.70) and patients with impaired liver function not due to Wilson’s disease (r = 0.65). Conclusions: For the enzymatic assay the best cut-off point for predicting Wilson’s disease was estimated to be 55 U/ L. Our data suggest that the enzymatic ceruloplasmin assay is superior to the immunologic assay in diagnosing Wilson’s disease and should become the preferred method. Ó 2009 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved. Keywords: Wilson’s disease; Ceruloplasmin; Oxidase activity

1. Introduction Wilson’s disease (WD) is a rare, autosomal-recessively inherited disorder of copper metabolism due to

Received 7 February 2009; received in revised form 27 May 2009; accepted 16 June 2009; available online 30 July 2009 Associate Editor: Y.M. Deugnier q The authors who have taken part in this study declared that they do not have anything to disclose regarding funding from industries or conflict of interest with respect to this manuscript. * Corresponding author. Tel.: +49 6221 5638709; fax: +49 6221 564116. E-mail address: [email protected] (U. Merle). Abbreviations: WD, Wilson’s disease; ROC, receiver operating characteristics; CI, confidence interval.

mutations of the WD gene ATP7B [1,2]. WD is clinically characterized by hepatic (e.g. liver cirrhosis) and neurological manifestations related to the accumulation of copper in the liver and other organs, e.g. the brain. Early diagnosis is essential because specific, life-long treatment can prevent further liver injury and neurological complications in most cases. Especially in patients with liver disease diagnosis of WD is often difficult as none of the commonly used diagnostic parameters gives a reasonable degree of certainty to establish or to exclude the diagnosis of WD [3–5]. The WD gene ATP7B codes for a membrane-bound, copper-binding protein that is expressed primarily in the liver [1,2]. Disease causing mutations in the ATP7B gene result in a reduced biliary copper excretion and an

0168-8278/$36.00 Ó 2009 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.jhep.2009.06.022

926

U. Merle et al. / Journal of Hepatology 51 (2009) 925–930

impaired hepatic holo-ceruloplasmin synthesis. The latter, however, seems to be of minor importance in copper homeostasis, since patients with aceruloplasminemia were shown to have almost normal copper supply to organs [6]. Although ceruloplasmin deficiency therefore might only represent an epiphenomenon of WD, serum ceruloplasmin concentration is considered a useful laboratory test for diagnosis of WD [4]. However, in a study by Steindl et al. [3] only 73% of mutation-confirmed WD patients, who presented with chronic liver disease had immunoreactive serum ceruloplasmin concentrations 1.5) due to liver cirrhosis (of an etiology other than WD). Of the 52 healthy subjects 20 (38.5%) were male. The median age of healthy controls was 34 years (IQR 30.5–42). Of the 51 patients with impaired liver function 30 (58.8%) were male. The median age was 56 years (IQR 49.5–61). All subjects were enrolled into the study by giving their written informed consent, and the study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki as reflected in a priori approval by the ethical committee of the Medical University of Heidelberg.

2.2. Laboratory methods Blood was collected after an overnight fast. Sera were collected and stored (for 4–28 months) at 80 °C until assayed. Total serum ceruloplasmin concentration was analyzed by the Department of clinical chemistry of the University hospital Heidelberg using an immuno-nephelometric assay with anti-serum to human ceruloplasmin (Dade Behring, Marburg, Germany) on BNII nephelometer (Dade Behring, Germany). The assay was used in line with the manufacturer’s instructions. Serum ceruloplasmin oxidase activity with o-dianisidine dihydrochloride as substrate was determined spectrophotometrically by one technician as described previously by Schosinsky et al. [12]. All measurements were performed in duplicates. All solutions used for measurement of ceruloplasmin oxidase activity were stored refrigerated at 4 °C until use (with o-dianisidine dihydrochloride solution stored light protected). Acetate buffer (pH 5, ionic strength 0.1) was prepared by mixing 990 ml ddH2O, 13.608 g sodium acetate trihydrate (analytical grade, Merck, Darmstadt, Germany), and 2.6 ml of glacial acetic acid (analytical grade, Baker Chemical Co., Deventer, The Netherlands), and subsequently adjusting pH to 5.0 by addition of 0.1 M glacial acetic acid. A volume of 0.75 ml of acetate buffer and 0.05 ml of serum was mixed and half pipetted into each of two reaction tubes. After placement and 5 min temperature equilibration in a 30 °C water bath, 0.1 ml of 7.88 mM o-dianisidine dihydrochloride (preincubated at 30 °C) (Sigma–Aldrich, Deisenhofen, Germany) was added at timed intervals in both tubes. At different time points (5 and 15 min) after adding o-dianisidine dihydrochloride both tubes were removed from the water bath. Directly after removing each tube 1.0 ml 9 M sulfuric acid (analytical grade, Baker Chemical Co., Deventer, The Netherlands) was added and solution mixed thoroughly. Next, absorbance of the purplish-red solutions was measured spectrophotometrically (Ultrospec III, Pharmacia, Cambridge, UK) at 540 nm vs. ddH2O as a blank. The enzymatic activity of ceruloplasmin was expressed in International Units, in terms of substrate consumed: Ceruloplasmin oxidase activity = (A15 A5) 6.25 102 U/L (for details of basis of calculation see [12]). Linearity of ceruloplasmin oxidase assay was assessed using two samples and one commercially available purified bovine ceruloplasmin preparation (Sigma–Aldrich, Deisenhofen, Germany). The samples and the commercially available ceruloplasmin preparation were assayed undiluted and serially diluted in ddH2O to obtain 50%, 25%, 12.5%, 6.25%, 3.125%, and 1.562% of the analyte. For intra-assay repeatability, four different samples were assayed 10 times on one day. For inter-assay repeatability, two samples and two dilutions of a commercially available purified ceruloplasmin preparation were assayed on 9 occasions over a period of 12 days. As the immunological ceruloplasmin assay and the enzymatic assay were performed in different laboratories by different persons, both were blind to the results of the other test.

2.3. Statistical analysis Variables are expressed as median with interquartile range (IQR). Intra- and inter-assay variation was described by the coefficient of variation. The data obtained for ceruloplasmin oxidase activity and ceru-

U. Merle et al. / Journal of Hepatology 51 (2009) 925–930 loplasmin concentration are summarized as box and whisker plots. For statistical analysis of differences between the groups the Kruskal–Wallis test was used. McNemar test was used to compare sensitivity and specificity of both ceruloplasmin assays. As recommended by Trajman and Luiz [14] sensitivity and specificity were evaluated separately by McNemar test – among WD patients and among controls, respectively. Correlational analyses were performed with Spearman rank order correlation and correlation coefficients are given. All calculations including generation of receiver operating characteristics (ROC) analysis were performed using SPSS version 16.0 (SPSS, Chicago, IL). Two sided P values were reported in all cases and a P < 0.05 was considered statistically significant.

3. Results 3.1. Validation of serum ceruloplasmin oxidase activity test Dilution linearity was observed in the range of 2.0– 250.0 U/L (with a regression coefficient of r2 > 0.999). The ranges of intra- and inter-assay coefficients of variation were 2.9–7.7% and 5.0–9.3%, respectively. 3.2. Comparison of serum ceruloplasmin oxidase activity with immunoreactive ceruloplasmin concentration Ceruloplasmin oxidase activity and immunoreactive ceruloplasmin concentration were determined in 110 consecutive WD patients, all referred to the Department

927

of Gastroenterology at the University of Heidelberg between 04/2006 and 04/2008. Both laboratory tests were performed on frozen sera after finishing collection of samples. Of the 110 WD patients analyzed, the primary clinical symptoms were neurological in 29 and hepatic in 71 patients, while 10 patients were asymptomatic. Serum ceruloplasmin oxidase activity was significantly lower in WD patients (6.9 U/L (IQR 1.9–21.3)) than in healthy controls (100.9 U/L (IQR 83.8–147.0), P < 0.0001) and non-WD patients with impaired liver function (80.6 U/ L (IQR 38.0–110.2), P < 0.0001), respectively (Fig. 1A). Serum ceruloplasmin concentration was also significantly lower in WD patients (0.10 g/L (IQR 0.06–0.14)) than in healthy controls (0.24 g/L (IQR 0.21–0.32), P < 0.0001) and non-WD patients with impaired liver function (0.21 g/L (IQR 0.17–0.25), P < 0.0001) (Fig. 1B). In the cohort of WD patients and healthy subjects the ROC curve analysis (Fig. 2A) resulted in the greatest sum of sensitivity and specificity of serum ceruloplasmin oxidase activity in identifying WD patients at a cut-off point of 55 U/L. With this cut-off (655 U/L) the sensitivity was 93.6% (95% CI, 86.9–97.2%), and the specificity was 100% (95% CI, 91.4–100%). By ROC curve analysis the greatest sum of sensitivity and specificity of immunoreactive ceruloplasmin concentrations in our study cohort was obtained at a cut-off point of

Fig. 1. Serum ceruloplasmin oxidase activity and immunoreactive serum ceruloplasmin in different groups of Wilson’s disease patients, healthy controls and non-Wilson’s disease patients with impaired liver function. Distribution of data presented as a box and whisker plots: 25th percentile, median, 75th percentile, maximum, outliers (), and extreme values (). Both assay resulted in no indeterminate results or missing data. (A) Serum ceruloplasmin oxidase activity in Wilson disease (WD) patients, healthy controls and non-WD patients with impaired liver function. Oxidase activity is significantly lower in WD patients (6.9 U/L (interquartile range (IQR) 1.9–21.3)) compared to healthy controls (100.9 U/L (IQR 83.8–147.0)) (P < 0.0001) and nonWD patients with impaired liver function (80.6 U/L (IQR 38.0–110.2)) (P < 0.0001). (B) Immunoreactive serum ceruloplasmin is significantly lower in WD patients (0.10 g/L (IQR 0.06–0.14)) compared to healthy controls (0.24 g/L (IQR 0.21–0.32)) (P < 0.0001) and non-WD patients with impaired liver function (0.21 g/L (IQR 0.17–0.25)) (P < 0.0001). [This figure appears in colour on the web.]

928


Fig. 2. ROC (receiver operating characteristic) curves for the prediction of Wilson’s disease by serum ceruloplasmin oxidase activity and immunoreactive serum ceruloplasmin. (A) When analyzing Wilson’s disease (WD) patients together with healthy controls for serum ceruloplasmin oxidase activity (bold line) the area under the curve is 0.984 (95% CI, 0.970–0.999) and for immunoreactive ceruloplasmin (light line) the area under the curve is 0.954 (95% CI, 0.926–0.983). (B) When analyzing WD patients together with healthy controls and non-WD patients with impaired liver function for serum ceruloplasmin oxidase activity (bold line) the area under the curve is 0.945 (95% CI, 0.917– 0.974) and for immunoreactive ceruloplasmin (light line) the area under the curve is 0.930 (95% CI, 0.897–0.962). Diagnosis of WD was made using the WD score published previously [13].

0.19 g/L (Fig. 2A). At this cut-off (60.19 g/L) the sensitivity was 93.6% (95% CI, 86.9–97.2%) and the specificity was 78.8% (95% CI, 64.9–88.5%). When additionally including the 51 non-WD patients with impaired liver function into ROC curve analysis (Fig. 2B) for identifying patients with WD the sensitivity of the enzymatic assay (cut-off 55 U/L) was 93.6% (95% CI, 86.9– 97.2%) and the specificity 84.5% (95% CI, 75.7–90.6%). The sensitivity and specificity of immunoreactive ceruloplasmin concentrations (cut-off 60.19 g/L) in this study cohort were 93.6% (95% CI, 86.9–97.2%) and 68.9% (95% CI, 58.9–77.5%), respectively. The ROC curve analysis showed that serum ceruloplasmin oxidase activity and immunoreactive serum ceruloplasmin show no difference in their sensitivities for diagnosing WD (P = 0.75, McNemar test). With regard to specificity the enzymatic assay (specificity 100%) was significantly

better than the immunologic assay (specificity 78.8%) when analyzing healthy controls (P = 0.02, McNemar test). When additionally including patients with impaired liver function, the specificity of the enzymatic assay was higher than that of the immunologic assay (78.8% vs. 68.9%). However, this difference did not reach statistical significance (P = 0.10, McNemar test). Analyzing the small subgroup of treatment-naı¨ve WD patients separately, all of these 14 patients had serum ceruloplasmin levels 60.19 g/L and 13 had ceruloplasmin oxidase activities 655 U/L. Healthy controls demonstrated a strong positive correlation between oxidase activity and ceruloplasmin protein levels (r = 0.94; 95% CI, 0.91–0.97) (Fig. 3A). In WD patients the correlation between oxidase activity and ceruloplasmin protein levels was weaker (r = 0.70; 95% CI, 0.6–0.79) (Fig. 3B). Of note, in non-WD patients with impaired liver function correlation strength of oxidase activity and ceruloplasmin protein level was also lower than in healthy controls (r = 0.65; 95% CI, 0.46–0.79) (Fig. 3C). Analysis of four heterozygous relatives of index WD patients revealed ceruloplasmin oxidase activities of 60.7, 89.7, 57.1, and 102.3 U/L, respectively, and immunoreactive ceruloplasmin concentrations of 0.19, 0.24, 0.21, and 0.28 g/L, respectively. However, these data have to be confirmed on a more significant series.

4. Discussion Because the establishment of a correct diagnosis before starting treatment is crucial in WD, valid diagnostic tests are of great importance. The established diagnostic parameters urinary copper excretion, serum ceruloplasmin concentration, and serum copper level are known to have sensitivities and specificities of only 80–90% [3,4,15]. Measurement of hepatic copper content has a high sensitivity (96.5% at a cut-off of >75 lg/g dry weight) [16], but requires an invasive liver biopsy. In our study, we could show that the enzymatic ceruloplasmin assay has a much higher specificity for diagnosing WD than the immunologic assay (100% vs. 78.8%), while both tests have the same sensitivities (93.6% for both). ROC curve analysis with a control cohort including patients with impaired liver function not due to WD revealed reduced specificities for both ceruloplasmin tests. Thus, in our opinion, ceruloplasmin oxidase activity is a useful diagnostic test for WD in patients with no severe impairment of liver function. Especially in patients where standard diagnostic tests give conflicting results, ceruloplasmin oxidase activity may provide additional important information. The enzymatic method measures the biological active holo-ceruloplasmin exclusively, while the immunologic ceruloplasmin test measures total ceruloplasmin [8]. Espe-


929

ous studies on measurement of ceruloplasmin oxidase activity in WD patients with p-phenylenediamine as substrate, the ceruloplasmin levels obtained with an immunological method were higher than those measured enzymatically [8,17]. The discrepancy of ceruloplasmin levels obtained by the enzymatic and the immunologic ceruloplasmin assay is reflected in our study in the weaker correlation of both parameters in WD patients when compared to correlation in healthy controls. Of interest, patients with impaired liver function not due to WD also displayed a diminished correlation of the immunologically and enzymatically measured ceruloplasmin values when compared to healthy controls. One might speculate that this might reflect an impaired copper incorporation into apo-ceruloplasmin in impaired liver function. Walshe [8] reported in 2003 and Macintyre et al. [17] in 2004 their experience with measurement of ceruloplasmin oxidase activity using p-phenylenediamine as substrate. Although both reports quoted the enzymatic measurement of ceruloplasmin as superior to the immunological method, to date the enzymatic method has not found its way into clinical practice [5]. A reason might be the need for special precautions when using p-phenylenediamine as substrate [18]. Reported inhibitors of p-phenylenediamine oxidase activity include ascorbic acid [18], citrate [19], and ethylenediaminetetraacetate [17,19]. In contrast, for o-dianisidine as substrate Schosinsky et al. [12] have already shown thato-dianisidine is more stable and less interference-prone than p-phenylenediamine and that there is little or no interference from reducing or colored components of serum. In our study we could show for the first time the usefulness of the determination of ceruloplasmin oxidase activity in WD patients using the method described by Schosinsky et al. [12]. In our opinion, the method is adequately simple to be suitable for routine clinical use. Our data show that measurement of serum ceruloplasmin oxidase activity is a discriminating biochemical test to diagnose or exclude WD non-invasively. In our study cohort, this test yielded an excellent specificity and high sensitivity for diagnosis of WD. Acknowledgements

Fig. 3. Scattergram of the immunochemically determined serum ceruloplasmin concentration plotted against the enzymatically determined serum ceruloplasmin oxidase activity. Analysis is performed in the cohorts of healthy controls (A), Wilson’s disease (WD) patients (B), and non-WD patients with impaired liver function (C). The estimated regression line and the correlation coefficient r is displayed for each cohort.

cially in WD patients total ceruloplasmin and the biological active form often differ a lot. As shown in two previ-

The authors thank Prof. P. Ferenci for performing ATP7B mutational analysis. The underlying research reported in the study was supported by grants to U.M. and W.S. from Deutsche Forschungsgemeinschaft and Dietmar Hopp Foundation. References [1] Bull PC, Thomas GR, Rommens JM, Forbes JR, Cox DW. The Wilson disease gene is a putative copper transporting P-type ATPase similar to the Menkes gene. Nat Genet 1993;5:327–337.

930


[2] Tanzi RE, Petrukhin K, Chernov I, Pellequer JL, Wasco W, Ross B, et al. The Wilson disease gene is a copper transporting ATPase with homology to the Menkes disease gene. Nat Genet 1993;5:344–350. [3] Steindl P, Ferenci P, Dienes HP, Grimm G, Pabinger I, Madl C, et al. Wilson’s disease in patients presenting with liver disease: a diagnostic challenge. Gastroenterology 1997;113:212–218. [4] Roberts EA, Schilsky ML. A practice guideline on Wilson disease. Hepatology 2003;37:1475–1492. [5] Roberts EA, Schilsky ML. Diagnosis and treatment of Wilson disease: an update. Hepatology 2008;47:2089–2111. [6] Harris ZL, Takahashi Y, Miyajima H, Serizawa M, MacGillivray RT, Gitlin JD. Aceruloplasminemia: molecular characterization of this disorder of iron metabolism. Proc Natl Acad Sci USA 1995;92:2539–2543. [7] Walshe JM, Briggs J. Caeruioplasmin in liver disease. A diagnostic pitfall. Lancet 1962;2:263–265. [8] Walshe JM. Wilson’s disease: the importance of measuring serum caeruloplasmin non-immunologically. Ann Clin Biochem 2003;40:115–121. [9] Kojimahara N, Nakabayashi H, Shikata T, Esumi M. Defective copper binding to apo-ceruloplasmin in a rat model and patients with Wilson’s disease. Liver 1995;15:135–142. [10] Ravin HA. An improved colorimetric enzymatic assay of ceruloplasmin. J Lab Clin Med 1961;58:161–168.

[11] Broman L. Separation and characterization of two coeruloplasmins from human serum. Nature 1958;182:1655–1657. [12] Schosinsky KH, Lehmann HP, Beeler MF. Measurement of ceruloplasmin from its oxidase activity in serum by use of odianisidine dihydrochloride. Clin Chem 1974;20:1556–1563. [13] Ferenci P, Caca K, Loudianos G, Mieli-Vergani G, Tanner S, Sternlieb I, et al. Diagnosis and phenotypic classification of Wilson disease. Liver Int 2003;23:139–142. [14] Trajman A, Luiz RR. McNemar. x2 test revisited: comparing sensitivity and specificity of diagnostic examinations. Scan J Clin Lab Invest 2008;68:77–80. [15] Perman JA, Werlin SL, Grand RJ, Watkins JB. Laboratory measures of copper metabolism in the differentiation of chronic active hepatitis and Wilson disease in children. J Pediatr 1979;94:564–568. [16] Ferenci P, Steindl-Munda P, Vogel W, Jessner W, Gschwantler M, Stauber R, et al. Diagnostic value of quantitative hepatic copper determination in patients with Wilson’s disease. Clin Gastroenterol Hepatol 2005;3:811–818. [17] Macintyre G, Gutfreund KS, Martin WR, Camicioli R, Cox DW. Value of an enzymatic assay for the determination of serum ceruloplasmin. J Lab Clin Med 2004;144:294–301. [18] Henry RC, Jacobs SL, Segalove M. Determination of ceruloplasmin oxidase in serum. Proc Soc Exp Med Biol Med 1960;104:620. [19] Osaki S, McDermott JA, Frieden E. Citric acid as the principal serum inhibitor of ceruloplasmin. J Biol Chem 1964;239:364–366.