Familial and Sporadic Porphyria Cutanea Tarda: Characterization and Diagnostic Strategies

Transcription

1 Papers in Press. Published February 20, 2009 as doi: /clinchem The latest version is at Clinical Chemistry 55: (2009) Evidence-Based Medicine and Test Utilization Familial and Sporadic Porphyria Cutanea Tarda: Characterization and Diagnostic Strategies Aasne K. Aarsand, 1* Helge Boman, 2,3 and Sverre Sandberg 1,4 BACKGROUND: Porphyria cutanea tarda (PCT) occurs in sporadic (spct) and familial (fpct) forms, which are generally clinically indistinguishable clinically and have traditionally been differentiated by erythrocyte uroporphyrinogen decarboxylase (UROD, EC ) activity. We used UROD gene sequencing as the reference in assessing the diagnostic accuracy of UROD activity and the mutation spectrum of the UROD gene, evaluating the frequency and disease attributes of PCT and its subtypes in Norway, and developing diagnostic models that use clinical and laboratory characteristics for differentiating fpct and spct. METHODS: All consecutive patients with PCT diagnosed within a 5-year period were used for incidence calculations. UROD activity analysis, UROD gene sequencing, analysis of hemochromatosis mutations, and registration of clinical and laboratory data were carried out for 253 patients. RESULTS: Fifty-three percent of the patients had disease-relevant mutations, 74% of which were c.578g Cor c.636 1G C. The UROD activity at the optimal cutoff had a likelihood ratio (LR) of 9.2 for fpct, whereas a positive family history had an LR of 19. A logistic regression model indicated that low UROD activity, a high uroporphyrin/heptaporphyrin ratio, a young age at diagnosis, male sex, and low alcohol consumption were predictors of fpct. The incidence of PCT was 1 in CONCLUSIONS: Two commonly occurring mutations are responsible for the high frequency of fpct in Norway. UROD activity has a high diagnostic accuracy for differentiating the 2 PCT types, and a model that takes into account both clinical information and laboratory test results can be used to predict fpct American Association for Clinical Chemistry Porphyria cutanea tarda (PCT) 5 is a disorder of porphyrin metabolism with associated skin photosensitivity that usually presents with vesiculobullous eruptions on the hands and face, and signs of liver damage. The disease is caused by a deficiency in uroporphyrinogen decarboxylase (UROD), the fifth enzyme in the heme synthesis pathway, and can be classified into several types, familial PCT (fpct) and sporadic PCT (spct) being the most common. fpct, an autosomal dominant disorder of low penetrance characterized by low UROD activity in all cells, constitutes about 25% of the cases in most populations (1 7). The level of UROD activity in erythrocytes has traditionally been used to distinguish between fpct and spct, because spct shows reduced UROD activity only in the liver. Both forms are considered clinically indistinguishable and are precipitated by the same factors: iron overload, infection by hepatitis C virus, and the use of estrogens and alcohol (8, 9). The increasing public interest in hereditary diseases and genetic testing is likely to make differentiation between spct and fpct increasingly important and consequently produce a greater demand for DNA analysis. In our experience, many PCT patients want to know whether they have a hereditary susceptibility for the disease, and their healthy family members want to be able to use this knowledge to consider precautions, particularly regarding the use of birth control pills in young women. The aims of this study were (a) to examine the diagnostic accuracy of UROD activity as a fpct marker, with UROD 6 gene sequencing as the reference standard; (b) to estimate fpct frequency and the spectrum of mutations responsible for fpct in a large, ethnically homogeneous 1 Norwegian Porphyria Centre (NAPOS), Laboratory of Clinical Biochemistry; and 2 Center for Medical Genetics and Molecular Medicine, Haukeland University Hospital, Bergen, Norway; 3 Section of Medical Genetics and Molecular Medicine, Department of Clinical Medicine, University of Bergen, Bergen, Norway; 4 Norwegian Quality Improvement of Primary Care Laboratories (NOKLUS), Section for General Practice, University of Bergen, Bergen, Norway. * Address correspondence to this author at: Norwegian Porphyria Centre (NAPOS), Laboratory of Clinical Biochemistry, Haukeland University Hospital, NO-5021 Bergen, Norway. Fax ; aasne.aarsand@helse-bergen.no. Received September 5, 2008; accepted January 13, Previously published online at DOI: /clinchem Nonstandard abbreviations: PCT, porphyria cutanea tarda; UROD, uroporphyrinogen decarboxylase; fpct, familial PCT; spct, sporadic PCT; uroheptaporphyrin ratio, uroporphyrin heptaporphyrin ratio; OR, odds ratio; LR, likelihood ratio. 6 Human genes: UROD, uroporphyrinogen decarboxylase; HFE, hemochromatosis. 1 Copyright (C) 2009 by The American Association for Clinical Chemistry

2 group of PCT patients; (c) to identify what other factors, including HFE (hemochromatosis) genotypes, characterize the 2 subtypes; (d) to use these factors and UROD activity to develop diagnostic models for differentiating spct and fpct; and (e) to estimate the incidence of PCT in Norway. Materials and Methods PARTICIPANTS We diagnosed PCT according to established criteria (8) and calculated disease incidence from all patients with PCT diagnosed at the Norwegian Porphyria Centre during and the first 5 months of 2006 (n 251). Material for DNA analysis was obtained from 196 of the patients and from 57 patients with previously diagnosed PCT who had control samples analyzed during the study period. These 253 patients were enrolled in the other parts of the study. Ostensibly healthy family members at risk for the disease (n 74) were tested during the study period and were included in parts of the study. The UROD gene was sequenced from PCT patients for whom DNA samples were available. Seven of these patients possessed sequence variant c.1104*3g A, which has previously been reported as a likely disease-causing mutation (10). This variant coincided with well-established UROD mutations in 2 of these patients. The biochemical and clinical presentations of these 2 patients, however, were not compatible with hepatoerythropoietic porphyria. This variant was also seen in 3 of 200 healthy control individuals and was therefore categorized as a single-nucleotide polymorphism in our material (Table 1, II). UROD sequence variants of undetermined significance were identified in 5 patients, and c.745c T coincided with c.876 7_878dup10 in one of these patients (Table 1, III). After excluding these 5 patients, we constituted the remaining 248 symptomatic index patients, hereafter denoted as symptomatic PCT patients, as the main study individuals in this study. We analyzed HFE mutants encoding the C282Y and H63D variants in all patients and conducted analyses of the HFE mutant encoding the S65C variant in 243 patients and analyses of UROD activity and reticulocytes in 246 patients. For the 74 healthy family members at risk, we carried out genetic and biochemical testing primarily as for the symptomatic patients. We collected material for enzyme and DNA analyses simultaneously for each participant and carried out porphyrin analyses at the time of diagnosis for the symptomatic patients and at the time of genetic testing for the healthy family members at risk. A standardized questionnaire filled out by each patient s personal physician was used to gather information regarding time of disease onset, symptoms, provoking factors, liver- and iron-related biochemical variables at diagnosis, maltreatment, and family history. Questionnaires were returned for 73% of the patients. The study was approved by the Regional Ethics Committee of Western Norway. BIOCHEMICAL METHODS We removed the buffy coat of heparin- or EDTA-anticoagulated blood and washed the erythrocytes in 0.15 mol/l NaCl. Erythrocyte UROD activity was measured essentially as previously described (11). Reticulocytes were enumerated on the Cell-Dyn 4000 (Abbott Diagnostics) in a subsample of the washed blood sample. Analysis of urinary -aminolevulinic acid and porphobilinogen with the ALA/PBG by Column Test (Bio-Rad Laboratories) was followed by spectrophotometric measurements. Plasma samples were fluorometrically scanned as previously described (12). Urine and fecal samples were dissolved in hydrochloric acid and phosphate buffer. Porphyrins were then extracted on a solid-phase column, separated by HPLC (BetaBasic Columns; Thermo Hypersil Limited) with a gradient of a solution of 1 g/l trifluoroacetic acid and acetonitrile (volume ratio, 70:30) against acetonitrile, and detected fluorometrically (excitation, 403 nm; emission, 618 nm). Urinary creatinine was assayed on the Hitachi Modular P instrument (Roche Diagnostics). The total analytical CV (within-day and between-day) for a UROD activity of 1.8 U/L erythrocytes was 5.2% during the study period. DNA SEQUENCING AND MUTATION DETECTION DNA was purified from EDTA-treated blood with the BioRobot M48 workstation (Qiagen) according to the manufacturer s instructions. We used Oligo 6.3 software (Molecular Biology Insights) to design PCR primers for amplification of exons and flanking introns (at least 50 bp) for the UROD and HFE (exon 2 for S65C) genes. PCR and real-time PCR amplifications were performed under standard conditions with AmpliTaq Gold (Applied Biosystems) or Taq polymerase (Qiagen). The PCR products were treated with shrimp alkaline phosphatase/exonuclease I (Amersham/GE Healthcare) and sequenced with the BigDye Terminator Cycle Sequencing Kit, version 2 (Applied Biosystems) according to the manufacturer s instructions. The HFE mutants encoding the H63D and C262Y variants were determined by real-time allelic discrimination with the 7900HT Fast Real-Time PCR System (Applied Biosystems) according to the manufacturer s instructions. The presence of the S65C variant was determined by sequencing HFE exon 2 or via restriction enzyme analysis with HinfI (New England BioLabs). DNA samples from most patients with c.578g C were tested for up to 43 microsatellite marker loci surround- 2 Clinical Chemistry 55:4 (2009)

3 Familial and Sporadic Porphyria Cutanea Tarda Table 1. DNA sequence variants in the UROD locus observed in Norwegian PCT patients (n 253). a Patients, n Sequence variants Protein prediction Population frequency Reference I. Likely causative mutations 5 c.20 1G A b Splice defect This study 3 c.59t C b L20P This study 3 c.100t A b W34R This study 5 c.238g T A80S Brady et al. (1) 2 c.448c T b P150S This study 4 c.568c T b Q190X This study 39 c.578g C R193P Phillips et al. (15) 58 c.636 1G C Splice defect Garey et al. (16) 3 c.850t C b W284R This study 7 c.876 7_878dup10 b p.? This study 1 c.910_912delaac b p.asn304del This study 1 c.1046_1047delat b p.his349argfsx10 This study II. Likely single-nucleotide polymorphisms c.20 79T G b c.21 12C G b,c c C T rs c.450g A P150P rs c G C rs c.603a G P201P rs c.758t A L253Q rs /200 d c.1104*3g A 3/200 d Cappellini et al. (10) c.1104*70g A rs III. Undetermined variants 3 c.745c T b R249W 0/200 d This study 1 c.758t C b L253P 0/200 d This study 1 c.578g A b R193H This study a Reference sequences NM_ and ENSG b Sequence variant not listed as a likely causative mutation in the Human Gene Mutation Database (26) or as a variant single-nucleotide polymorphism in Ensembl (27). c Reverse-transcription PCR analysis showed transcripts of typical size only. d Number of heterozygotes observed among 200 healthy control individuals. ing the UROD locus on chromosome 1. Diagnostic samples from a few relatives of index patients were available for establishing definite haplotypes. STATISTICAL ANALYSIS Analyse-it (Analyse-it Software) and Microsoft Excel in Microsoft Office 2003 were used in ROC curve analysis to evaluate the diagnostic accuracy of UROD activity. General statistical analysis, group-comparison testing, table analysis, logistic regression, and correlation studies were carried out with SPSS 15.0 for Windows (SPSS). Logistic regression analysis (with forward stepwise variable selection) was performed with the PCT disease type (sporadic/familial) as the dependent variable and the following sets of independent variables: (a) UROD activity alone (n 246); (b) UROD activity, HFE status with the genotypes in Table 2 and the wild-type/wild-type genotype as the reference category, sex (male, 0; female, 1), age at diagnosis, uroporphyrin heptaporphyrin ratio (uroheptaporphyrin ratio), ferritin, -glutamyltransferase, alanine aminotransferase, liver disease (absent, 0; present, 1), and high alcohol consumption [absent (0) or present (1), as scored by the patient s personal physician as a provok- Clinical Chemistry 55:4 (2009) 3

4 Table 2. HFE genotypes for symptomatic PCT patients (n 243) by spct and fpct subgroups and compared with healthy Norwegian blood donors. a C282Y H63D S65C All PCT patients (n 243), % b spct patients (n 113), % fpct patients (n 130), % General population (n 204), % / / / 6.6** / / / / / / 8.6*** / / / / / / 5.8* / / / / / / / / / / / / 38.7*** a Analysis of S65C was not performed in 5 PCT patients, and results for these patients are not included. b Indicated are frequencies significantly different from those of healthy Norwegians (*P 0.02; **P 0.001; ***P ). ing factor for PCT] (n 151); and (c) all variables in (b) except for UROD activity. Logistic regression analysis was also performed in the fpct and spct subgroups with age at symptom onset ( 50 years and 50 years) as the dependent variable and HFE status, sex, liver disease, high alcohol consumption, and estrogen use (absent, 0; present, 1) as the independent variables (n 166). CIs for indices of diagnostic accuracy were calculated by the Wilson score method (13). All P values are the results of 2-tailed tests, and P values 0.05 are considered statistically significant. Results PCT was diagnosed in 251 consecutive patients with an even distribution over the study period. The Norwegian Porphyria Centre performs porphyria diagnostics for the entire population, and this result yields a minimum estimate for the incidence of symptomatic PCT in Norway of 1.0 per inhabitants. UROD AND HFE GENOTYPES A disease-relevant sequence variant was identified in 131 (53%) of the 248 symptomatic PCT patients. UROD variants c.578g C and c.636 1G C accounted for 30% and 44% of the fpct cases, respectively (Table 1). There were no significant differences between sporadic and familial cases with regard to HFE status (Table 2). PATIENT CHARACTERISTICS Dividing the symptomatic PCT patients into sporadic and familial cases according to the results of UROD gene sequencing revealed the differences between the 2 groups described in Table 3), as well as additional differences. Alcohol was more often a provoking factor among men than among women (43% vs 25%), and liver disease, primarily infection with hepatitis B and C viruses, was more frequent in males with spct (29% vs 4%). Twenty-five percent of fpct patients reported maltreatment (mainly antibiotics, cortisone, or a combination of both), compared with 13% of the spct patients. fpct patients homozygous for the C282Y variant had a younger age of symptom onset (33 years) than either wild-type homozygotes (50 years) or all other HFE genotypes (49 years). A logistic regression analysis indicated that being male and the presence of liver disease predicted a younger age of disease onset in fpct [odds ratio (OR), 2.9; 95% CI, ] and spct (OR, 0.033; 95% CI, ), respectively. DIAGNOSTICS The reticulocyte percentage (range, 0% 4%) was not correlated with UROD activity. Fig. 1 shows the distribution of UROD activities in fpct and spct cases. Because the sample size was 200 (14), we used ROC curve analysis directly to identify the optimal cutoff of UROD activity. The cutoff of 1.3 U/L erythrocytes had a sensitivity of 96% (95% CI, 90% 99%), a specificity of 90% (95% CI, 82% 94%), a likelihood ratio (LR) of 9.2 (95% CI, ), and positive and negative predictive values of 91% (95% CI, 84% 95%) and 95% (95% CI, 90% 98%), respectively (Fig. 2). UROD activities and concentrations of urinary porphyrin metabolites were similar in the 3 major sequence-variant groups (i.e., c.578g C, c.636 1G C, and others ). Use of an uroheptaporphyrin ratio 2.6 as a marker for fpct had a sensitivity of 78% (95% CI, 70% 85%), a specificity of 51% (95% CI, 41% 60%), an LR of Clinical Chemistry 55:4 (2009)

5 Familial and Sporadic Porphyria Cutanea Tarda Table 3. Biochemical and clinical characteristics of fpct and spct patients in symptomatic PCT cases (n 248). a Biochemical or clinical variable spct fpct Male sex, % 45 (36 54) 57 (49 65) Age at first symptoms, years 53 (51 56)** 48 (45 51) Age at diagnosis, years 57 (55 59)*** 51 (48 53) UROD activity, U/L erythrocytes 1.7 ( )*** 0.9 ( ) Total porphyrins, nmol/mmol creatinine 883 (323, 2044) 901 (319, 2066) Uroporphyrins, nmol/mmol creatinine 556 (204, 1250) 673 (222, 1471) Heptaporphyrins, nmol/mmol creatinine 222 (77, 458) 198 (63, 471) Uroheptaporphyrin ratio 2.7 (1.6, 3.9)*** 3.2 (2.4, 5.8) Ferritin, g/l b 351 (148, 901)** 256 (113, 821) ALT, c U/L d 71 (39, 137)** 86 (49, 173) -GT, U/L e 125 (47, 329)*** 65 (31, 179) Vesicles on hands, % 92 (86 98) 88 (81 95) Vesicles on face, % 16 (8 24) 21 (12 30) Hyperpigmentation, % 30 (20 40) 31 (21 41) Hypertrichosis, % 29 (19 39) 27 (18 36) High alcohol consumption, % 48 (38 58)*** 22 (14 30) Use of estrogen, % 23 (14 32) 13 (6 20) Liver disease, % 16 (8 24)* 5 (1 9) Relatives with known PCT, % 2 (0 5)*** 32 (24 40) a The minimum number of patients with available data for clinical descriptors and liver variables was 159. Data are presented as the mean (95% CI) or as the median (10th percentile, 90th percentile), as indicated. For variables with nongaussian distributions, Student t-tests were performed after the data were transformed to natural logarithms. Indicated are statistically significant differences for spct vs fpct (*P 0.05; **P 0.01; ***P ). b Upper reference limits are 240 g/l for females and 300 g/l for males. c ALT, alanine aminotransferase; -GT, -glutamyltransferase. d Upper reference limits are 45 U/L for females and 70 U/L for males. e Upper reference limits for individuals older than 40 years are 75 U/L for females and 115 U/L for males. (95% CI, ), and positive and negative predictive values of 63% (95% CI, 55% 71%) and 68% (95% CI, 60% 75%), respectively (Fig. 2). Having a positive family history of PCT had an LR of 19 (95% CI, 5 76) for diagnosing fpct, and a negative or unknown family history had an LR of 0.7 (95% CI, ; Table 2). A logistic regression analysis with diagnosis of fpct vs spct as the dependent variable identified the following as significant discriminators: UROD activity (OR, ; 95% CI, ), uroheptaporphyrin ratio (OR, 2.94; 95% CI, ), age at diagnosis (OR, 0.91; 95% CI, ), sex (OR, 0.22; 95% CI, ), and high alcohol consumption (OR, 0.09; 95% CI, ). This model correctly predicted 93% of the cases. In some diagnostics laboratories, measurement of UROD activity is not available. In this setting, logistic regression analysis correctly classified 83% of the cases with the following significant variables: sex (OR, 0.23; 95% CI, ), age at diagnosis (OR, 0.96; 95% CI, ), uroheptaporphyrin ratio (OR, 3.34; 95% CI, ), ferritin (OR, 0.998; 95% CI, ), -glutamyltransferase (OR, 0.985; 95% CI, ), alanine aminotransferase (OR, 1.02; 95% CI, ), and high alcohol consumption (OR, 0.22; 95% CI, ). All models gave essentially the same ORs for the different discriminators when they were applied to the patient subgroups with and without a positive family history of PCT. PREDICTIVELY TESTED FAMILY MEMBERS Forty-three of the 74 tested healthy family members at risk for the disease carried their family s mutation. Eight of these 43 individuals were in a subclinical disease state and fulfilled the biochemical criteria for PCT. All urinary metabolites with the exception of coproporphyrins were significantly lower in the subclinical patients than in the symptomatic PCT patients (median uroporphyrin concentration, 140 and 605 nmol/ mmol creatinine, respectively). Clinical Chemistry 55:4 (2009) 5

6 UROD activity (U/L erythrocytes) Discussion spct (n = 115) fpct (n = 131) Fig. 1. UROD activity for symptomatic spct and fpct patients. Data are presented as the median and 25th and 75th quartiles. PCT is the most prevalent porphyria in most populations and differs from the other hereditary porphyrias in that it also has a sporadic form. Our study is the first to report that the familial form may constitute more than half of the cases in a large national sample ( 240 index patients). This frequency is significantly higher than frequencies reported for comparable populations Sensitivity specificity UROD activity Uroheptaporphyrin ratio Fig. 2. ROC curves for diagnostic performance of UROD activity and the uroheptaporphyrin ratio in identifying symptomatic patients with fpct. Areas under the ROC curve are 0.94 (95% CI, ) for UROD activity and 0.73 (95% CI, ) for the uroheptaporphyrin ratio. (1, 3, 5). The largest of these studies (1) identified 19 fpct cases in 84 patients, but because DNA analysis was performed only in patients with low UROD activity, the frequency of fpct might have been underestimated by a few percent. The 2 other studies reported the familial form to constitute about 25% of cases when UROD gene analysis was performed in all patients [53 and 61 patients, respectively (3, 5)], whereas 9 of 18 PCT patients in a Chilean study had familial disease (6). Most (74%) of the fpct cases in our study were caused by 2 frequently occurring mutations, c.578g C and c.636 1G C. Gene-marker studies of c.578g C indicate that this variant is a founder mutation that originated in the northwestern part of southern Norway (data not shown). c.578g C was first reported in a patient in Utah (15), and genemarker analysis demonstrated that this patient was a descendent of the same ancestor. The sequence variant c.636 1G C, which has been found in several unrelated patients, represents either an ancient mutational event or a hot spot for recurrent mutation (3, 15, 16). All of our patients with the c.636 1G C variant also carried the T allele in the single-nucleotide polymorphism c G T (rs ), an observation compatible with a common origin of this mutation on a UROD haplotype with the rarer allele (T) in position c Founder mutations have not previously been reported in PCT, in contrast to, for example, W198X in acute intermittent porphyria (northern Sweden) (17) and the R59W variant in variegate porphyria (South Africa) (18). We documented the incidence of PCT in Norway to be 1.0 per inhabitants. Excluding patients with c.578g C and c.636 1G C reduces the estimated incidence to 0.6 per Few reports have documented the incidence of PCT, but an estimate of per population has been reported for the UK (9). A recent metaanalysis has shown that HFE genotypes encoding variants C282Y and/or H63D are associated with a 2- to 48-fold increased risk of PCT compared with the wild-type homozygote, with the C282Y/ C282Y homozygote conferring the highest risk (19). Increased frequencies of C282Y homozygotes, C282Y/ H63D compound heterozygotes, and H63D homozygotes were seen in our PCT patients, compared with the general Norwegian population, but no association was seen for S65C. The percentage of C282Y homozygotes was only 6.6%, which is low compared with reports of 15% 25% in PCT patients from Britain and Scandinavia, in which the general population frequencies of HFE genotypes are comparable to those of Norway (1, 20, 21). The distributions of HFE genotypes in the spct and fpct patient groups were not significantly different in our study, contrary to what has been found 6 Clinical Chemistry 55:4 (2009)

7 Familial and Sporadic Porphyria Cutanea Tarda A Posttest probability of fpct B Posttest probability of fpct UROD activity (U/L erythrocytes) UROD activity (U/L erythrocytes) 20% pretest probability of fpct 50% pretest probability of fpct 90% pretest probability of fpct Female with high alcohol consumption Female with normal alcohol consumption Male with normal alcohol consumption Fig. 3. Posttest probabilities of fpct at different UROD activity concentrations. (A), UROD activity as the sole independent variable in populations with 20%, 50%, and 90% prevalence of fpct. (B), Consideration of sex and alcohol as a provoking factor in a 54-year-old patient with a uroheptaporphyrin ratio of in studies in which the 2 disease forms were differentiated on the basis of UROD activity (2, 21). In general, PCT, particularly the sporadic type, has been considered to occur more frequently in men (9, 22). In our study, however, 57% of the familial cases and 45% of the sporadic cases occurred in men (Table 2). With the exception of the uroheptaporphyrin ratio, the 2 groups showed no significant differences with respect to the urinary excretion of porphyrins, in contrast with the finding in a smaller study of a higher concentration of total urinary porphyrins in fpct patients (20). C282Y homozygosity has been reported to be associated with an earlier onset of symptoms, in both the familial and sporadic disease types (1). Our logistic regression analysis, however, showed that only liver disease and sex predicted a younger age of disease onset, in the sporadic and the familial groups, respectively. The consequence of demonstrating the presence of an HFE mutation in a patient with PCT is in itself not apparent, because PCT patients generally exhibit biochemical signs of iron overload. One study found that chloroquine is not as effective in C282Y homozygotes as in the other HFE genotypes, but the treatment of choice will be venesection in most cases, irrespective of the results of HFE testing (23). UROD activity differentiates sporadic and familial cases of PCT quite well (LR of 9.2 at 1.3 U/L erythrocytes. In comparison, porphobilinogen deaminase activity has been shown to differentiate between patients with acute intermittent porphyria and their healthy relatives with an LR of 3.6 (24). In several of the porphyrias, particularly erythropoietic protoporphyria, a small percentage of the patients with clear clinical and biochemical disease have no identifiable molecular defect (25). Not finding a molecular defect in a true fpct case in our study would have affected the diagnostic accuracy of UROD activity. Having one or more family members with overt PCT is a strong predictor of the familial type. In our study group with a 53% prevalence of the familial type, such a family history would give a posttest probability of fpct of 95% (Table 3), and a prevalence of 20% would give a posttest probability of about 80%. Not having or not knowing about relatives with PCT would reduce the posttest probability of fpct from 53% to 44% and reduce the prevalence from 20% to 15%. Posttest probabilities of fpct can also be calculated with various logistic regression models with different independent variables (Fig. 3). With the use of UROD activity as the sole independent variable and a pretest probability of fpct of 53%, a UROD result of 0.8 U/L erythrocytes will confirm fpct, and a result of 1.6 U/L will exclude fpct. Both kinds of results produce posttest probabilities of 90% (Fig. 3A). Applying the same model in a patient with a positive family history and a UROD result of 1.3 U/L will produce only minor increases in the high posttest probability of fpct. In a patient with a negative or unknown family history, however, a UROD result of 0.8 U/L increases the probability of fpct from 44% to 90%. Extending the diagnostic model to include clinical and biochemical information, such as the patient being a young man at the time of diagnosis with low alcohol consumption and a high uroheptaporphyrin ratio, further increases the probability of fpct (Fig. 3B). When this diagnostic model was used in the 5 patients excluded because of UROD sequence variants of undetermined significance (Table 1, III), the posttest probability of fpct in the patient who carried both the disease-causing c.876 7_878dup10 variant and c.745c T was 99%. The probability of fpct in the 4 other patients ranged from 7% to 34%. If measurement of UROD activity is not available, a diagnostic model incorporating sex, age at Clinical Chemistry 55:4 (2009) 7

8 diagnosis, uroheptaporphyrin ratio, ferritin, -glutamyltransferase, alanine aminotransferase, and alcohol consumption can be used to differentiate spct and fpct, but the model has low predictive power. The clinical value of discriminating fpct from spct is not clear. Although distinguishing the disease form does not appear to affect either the prognosis or the choice of treatment, our experience is that it is important for both patients and some of their family members to know whether the disease is hereditary. It is possible that the awareness of the hereditary aspect is higher in Norwegian patients, at least 20% of whom know of relatives with overt PCT. In addition, 20% of the healthy family members at risk in our study who had been demonstrated to carry their family s mutation had biochemically diagnosed PCT, indicating that the penetrance of fpct is possibly higher than has heretofore been assumed. Future follow-up of these and similar patients is therefore important to further elucidate the importance of differentiating between fpct and spct. A diagnostic strategy for differentiating fpct and spct can be derived from the information in this study, depending on what probability is deemed necessary for diagnosis, what clinical and laboratory data are known for the patient, what laboratory methods are References available, and how they have been standardized. When a diagnostic strategy is to be implemented, it is important to have an estimate of the local prevalence of fpct and to consider a possible spectrum bias. Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article. Authors Disclosures of Potential Conflicts of Interest: No authors declared any potential conflicts of interest. Role of Sponsor: The funding organizations played no role in the design of study, choice of enrolled patients, review and interpretation of data, or preparation or approval of manuscript. Acknowledgments: Data on UROD activity and UROD gene sequencing results for 71 of the patients described in this report were presented as a poster at the International Congress of Clinical Chemistry and Laboratory Medicine, Kyoto, Japan, We thank Solveig Blaflaat and Hanne Margrethe Jacobsen on behalf of the Norwegian Porphyria Centre and the Centre for Medical Genetics and Molecular Medicine, Haukeland University Hospital; Vo Tri Toan, University of Bergen, for his work on S65C in healthy Norwegians; and Dr. John D. Philips, University of Utah School of Medicine, for supplying a DNA sample from the previously described c.578g C patient (15). 1. Brady JJ, Jackson HA, Roberts AG, Whatley SD, Rowlands GL, Darby C, et al. Co-inheritance of mutations in the uroporphyrinogen decarboxylase and hemochromatosis genes accelerates the onset of porphyria cutanea tarda. J Invest Dermatol 2000;115: Bulaj ZJ, Phillips JD, Ajioka RS, Franklin MR, Griffen LM, Guinee DJ, et al. Hemochromatosis genes and other factors contributing to the pathogenesis of porphyria cutanea tarda. Blood 2000;95: Christiansen L, Bygum A, Jensen A, Brandrup F, Thomsen K, Horder M, Petersen NE. Uroporphyrinogen decarboxylase gene mutations in Danish patients with porphyria cutanea tarda. Scand J Clin Lab Invest 2000;60: Koszo F, Morvay M, Dobozy A, Simon N. Erythrocyte uroporphyrinogen decarboxylase activity in 80 unrelated patients with porphyria cutanea tarda. Br J Dermatol 1992;126: Mendez M, Poblete-Gutierrez P, Garcia-Bravo M, Wiederholt T, Moran-Jimenez MJ, Merk HF, et al. Molecular heterogeneity of familial porphyria cutanea tarda in Spain: characterization of 10 novel mutations in the UROD gene. Br J Dermatol 2007; 157: Poblete-Gutierrez P, Mendez M, Wiederholt T, Merk HF, Fontanellas A, Wolff C, Frank J. The molecular basis of porphyria cutanea tarda in Chile: identification and functional characterization of mutations in the uroporphyrinogen decarboxylase gene. Exp Dermatol 2004;13: Tavazzi D, Di Montemuros FM, Fargion S, Fracanzani AL, Fiorelli G, Cappellini MD. Levels of uroporphyrinogen decarboxylase (URO-D) in erythrocytes of Italian porphyria cutanea tarda patients. Cell Mol Biol (Noisy-le-grand) 2002;48: Deacon AC, Whatley SD, Elder GH. Porphyrins and disorders of porphyrin metabolism. In: Burtis CA, Ashwood ER, Bruns DE, eds. Tietz textbook of clinical chemistry and molecular diagnostics. 4th ed. New York: Elsevier Saunders; p Elder GH. Porphyria cutanea tarda. Semin Liver Dis 1998;18: Cappellini MD, Martinez di Montemuros F, Tavazzi D, Fargion S, Pizzuti A, Comino A, et al. Seven novel point mutations in the uroporphyrinogen decarboxylase (UROD) gene in patients with familial porphyria cutanea tarda (f-pct). Hum Mutat 2001;17: McManus J, Blake D, Ratnaike S. An assay of uroporphyrinogen decarboxylase in erythrocytes. Clin Chem 1988;34: Da Silva V, Simonin S, Deybach JC, Puy H, Nordmann Y. Variegate porphyria: diagnostic value of fluorometric scanning of plasma porphyrins. Clin Chim Acta 1995;238: Uitenbroek GD. SISA-Binomial (1997). quantitativeskills.com/sisa/statistics/diagnos.htm (Accessed July 2008). 14. Leeflang MM, Moons KG, Reitsma JB, Zwinderman AH. Bias in sensitivity and specificity caused by data-driven selection of optimal cutoff values: mechanisms, magnitude, and solutions. Clin Chem 2008;54: Phillips JD, Parker TL, Schubert HL, Whitby FG, Hill CP, Kushner JP. Functional consequences of naturally occurring mutations in human uroporphyrinogen decarboxylase. Blood 2001;98: Garey JR, Harrison LM, Franklin KF, Metcalf KM, Radisky ES, Kushner JP. Uroporphyrinogen decarboxylase: a splice site mutation causes the deletion of exon 6 in multiple families with porphyria cutanea tarda. J Clin Invest 1990;86: Andersson C. Acute intermittent porphyria in northern Sweden. A population-based study [Thesis]. Umeå, Sweden: Umeå Universitet; 1997: Meissner PN, Dailey TA, Hift RJ, Ziman M, Corrigall AV, Roberts AG, et al. A R59W mutation in human protoporphyrinogen oxidase results in decreased enzyme activity and is prevalent in South Africans with variegate porphyria. Nat Genet 1996;13: Ellervik C, Birgens H, Tybjaerg-Hansen A, Nordestgaard BG. Hemochromatosis genotypes and risk of 31 disease endpoints: meta-analyses including 66,000 cases and 226,000 controls. Hepatology 2007;46: Bygum A, Christiansen L, Petersen NE, Hørder M, Thomsen K, Brandrup F. Familial and sporadic porphyria cutanea tarda: clinical, biochemical and genetic features with emphasis on iron status. Acta Derm Venereol 2003;83: Harper P, Floderus Y, Holmström P, Eggertsen G, Gåfvels M. Enrichment of HFE mutations in Swed- 8 Clinical Chemistry 55:4 (2009)

9 Familial and Sporadic Porphyria Cutanea Tarda ish patients with familial and sporadic form of porphyria cutanea tarda. J Intern Med 2004;255: Anderson KE, Sassa S, Bishop DF, Desnick RJ. Disorders of heme biosynthesis: X-linked sideroblastic anemia and the porphyrias. In: Scriver CR, Beaudet AL, Valle D, Sly WS, Childs B, Kinzler KW, Vogelstein B, eds. The metabolic and molecular bases of inherited disease. 8th ed. New York: McGraw-Hill; p Stolzel U, Köstler E, Schuppan D, Richter M, Wollina U, Doss MO, et al. Hemochromatosis (HFE) gene mutations and response to chloroquine in porphyria cutanea tarda. Arch Dermatol 2003;139: Kauppinen R, von und zu Fraunberg M. Molecular and biochemical studies of acute intermittent porphyria in 196 patients and their families. Clin Chem 2002;48: Whatley SD, Mason NG, Holme SA, Anstey AV, Elder GH, Badminton MN. Gene dosage analysis identifies large deletions of the FECH gene in 10% of families with erythropoietic protoporphyria. J Invest Dermatol 2007;127: BIOBASE Web site. HGMD professional release (Accessed July 2008). 27. Ensembl Web site. (Accessed July 2008). Clinical Chemistry 55:4 (2009) 9