Familial Mediterranean fever in Germany | Springer Link

Download PDF
5 downloads 414 Views 310KB Size Report
Comment
Aug 19, 2012 - The origin of the disease appears to be the Middle East, but it has spread by migration all over the world. Sephardic Jews,. Armenians, North ...
Eur J Pediatr (2012) 171:1775–1785 DOI 10.1007/s00431-012-1803-8

ORIGINAL ARTICLE

Familial Mediterranean fever in Germany: epidemiological, clinical, and genetic characteristics of a pediatric population E. Lainka & M. Bielak & P. Lohse & C. Timmann & S. Stojanov & R. von Kries & T. Niehues & U. Neudorf

Received: 7 May 2012 / Revised: 9 July 2012 / Accepted: 16 July 2012 / Published online: 19 August 2012 # Springer-Verlag 2012

Abstract Familial Mediterranean fever (FMF) is an autoinflammatory disease and belongs to the heterogeneous group of hereditary recurrent fever syndromes (HRFs). Aims: The aims of the study were to determine the incidence of FMF in Germany and to describe the spectrum of pyrin mutations and the clinical characteristics in children. A prospective surveillance of children with HRF including FMF was conducted in Germany during a time period of 3 years by the German paediatric surveillance unit for rare paediatric diseases (ESPED). Monthly inquiries were sent to 370 children’s hospitals (Clinic-ESPED, n1) and to 23 laboratories (Laboratory-ESPED, n2). Inclusion criteria were children ≤16 years of age, disease-associated pyrin mutations, and more than three self-limiting episodes of fever >38.5 °C with increased inflammation markers. In n1, 122 patients with FMF and 225 pyrin mutations were identified. Ninety-two of 122 (75 %) children were of Turkish origin. The minimum incidence of FMF was estimated to be 3 (95 % CI: 2.48–3.54) E. Lainka (*) : M. Bielak : U. Neudorf Department of Paediatric Rheumatology, Children’s Hospital, University Duisburg-Essen, Hufelandstr. 55, 45122 Essen, Germany e-mail: [email protected] P. Lohse Department of Clinical Chemistry—Großhadern, University of Munich, Munich, Germany

per 106 person-years in the whole children population and 55 (95 % CI: 46–66) per 106 person-years in Turkish children living in Germany. N1 U n2 amounted to 593 asymptomatic and symptomatic carriers of 895 mutations (overlap of 73 cases with 134 mutations). p.Met694Val (45 %), p.Met680Ile (14 %), p.Val726Ala (12 %), and p.Glu148Gln (11.5 %) were the most common pyrin mutations. Conclusions: Despite FMF being the most frequent of the HRFs, its incidence in Germany is low. Twenty-five to 50 FMF patients ≤16 years are newly diagnosed per year. The disease is most commonly observed in individuals of Turkish ancestry. Key Messages • The incidence of FMF in Germany is calculated as 3 per 106 person-years for the entire children population and as 55 per 106 person-years for children of Turkish ancestry. • The pyrin p.Met694Val, p.Met680Ile, p.Val726Ala, and p.Glu148Gln mutations are the prevailing missense mutations. R. von Kries Institute of Social Paediatrics and Adolescent Medicine, University of Munich, Munich, Germany

T. Niehues Department of Paediatrics, HELIOS Hospital Krefeld, Krefeld, Germany

C. Timmann Department of Molecular Medicine, Bernhard Nocht Institute for Tropical Medicine, Hamburg, Germany S. Stojanov Department of Infectious Diseases and Immunology, Children’s Hospital, University of Munich, Munich, Germany

Present Address: P. Lohse Molecular Genetics Laboratory, Institute for Laboratory Medicine and Human Genetics, Singen, Germany

1776

Eur J Pediatr (2012) 171:1775–1785

• The awareness of the disease symptoms, especially for children with a migration background, must be increased, and the management of FMF as the most common AID must be improved.

analyze the clinical and genetic spectrum in a pediatric population-based sample using two independent methods for case ascertainment.

Keywords AID . FMF . MEFV gene . Pyrin . HRF . ESPED

Methods

Abbreviations FMF Familial Mediterranean fever ESPED German paediatric surveillance unit for rare paediatric diseases AID Autoinflammatory disease HRF Hereditary recurrent fever syndrome NSAID Non-steroidal anti-inflammatory drug MEFV Mediterranean fever

Study design

Introduction Familial Mediterranean fever (FMF, OMIM 249100) is an autoinflammatory disease (AID) associated with mutations in the Mediterranean Fever (MEFV, OMIM 608107) gene which is located on chromosome 16p13 and encodes the protein pyrin (formerly called marenostrin) [32, 33]. Pyrin is thought to play an essential role in a complex cellular system of interacting proteins, the inflammasome, which regulates inflammation induced by various infectious as well as physical and chemical stimuli in granulocytes [14]. Most FMF-associated mutations are single amino acid substitutions [8]. Spontaneously, resolving episodes of fever of 12- to 72-h duration, a serositis in form of peritoneal, pleural, or synovial inflammation, and erysipelaslike skin lesions, as well as increased acute phase reactants are typical features of the disease [28]. There is a risk of amyloidosis and a risk of unnecessary abdominal surgery, the latter caused by abdominal symptoms of FMF mimicking the clinical picture of an acute abdomen [3]. Colchicine is an established therapeutic and prophylactic regimen, preventing especially renal amyloidosis [2, 12, 41]. However, 5–10 % of the patients do not respond to colchicine [15, 22]. FMF is diagnosed mainly by clinical criteria. Genetic analysis is used to confirm the diagnosis and is especially helpful in oligosymptomatic patients [23]. Homozygous, compound heterozygous, and heterozygous mutations can be found in FMF patients [7]. The diagnosis is strongly suggested by patient ethnicity. The origin of the disease appears to be the Middle East, but it has spread by migration all over the world. Sephardic Jews, Armenians, North Africans, Turks, and to a lesser extent, Ashkenazi Jews, Greeks, and Italians are potential carriers [5]. However, individuals outside these groups were also shown to be affected [36]. The epidemiology and clinical manifestation of FMF as well as the spectrum of pyrin mutations in Germany have not been investigated until now [16]. Our aim was to estimate the incidence of this disorder in Germany and to

A prospective, national, active surveillance of symptomatic children with hereditary recurrent fever syndromes (HRF) was conducted by the German Paediatric Surveillance Unit for rare paediatric diseases (ESPED) from July 2003 until June 2006, as described previously [21]. Monthly inquiries were sent to 370 children’s hospitals (Clinic-ESPED, n1). As a second source of information, 23 laboratories performing genetic analyses for HRF were contacted monthly (Laboratory-ESPED, n2). Case definition The criteria to include patients were: age ≤16 years, confirmed MEFV gene mutation(s), and more than three selflimiting episodes of fever >38.5 °C of unknown origin associated with increased inflammation markers. Newly diagnosed patients with at least one mutation in the MEFV gene were added to the database, and epidemiological, clinical, and genetic informations were collected using questionnaires for hospitals and laboratories. The return rates were 97 % for the monthly report cards of Clinic-ESPED and above 90 % and 98 %, respectively, for the questionnaires of Clinic-ESPED and LaboratoryESPED. For each patient, the following data were documented in the Clinic-ESPED (n1): unique identification number, number and town of reporting clinic, core data (gender, date of birth (month/year), and time of diagnosis (month/ year)), history (consanguinity, ethnic origin, and affected relatives), symptoms (description of fever episodes, skin, joint, abdominal, neurological, and bone involvement, lymphadenopathy, serositis, amyloidosis, trigger factors, and rare observations), inflammation markers, genetic analysis, treatment (NSAIDs, corticosteroids, colchicine, etanercept, or other treatment); and in the LaboratoryESPED (n2): core data (see above), reporting institution (children’s hospitals or pediatric private practices), patient or relative, genetic analysis, and laboratory method. Consecutive patients with the tentative diagnosis of FMF were subjected to molecular genetic testing. Nine laboratories examined the MEFV gene in two steps: (1) exons 2, 3, and 10 (n02), or exons 2 and 10 (n03), or exons 2, 3, 5, and 10 (n01), or exons 1, 2, 3, 5, and 10 (n01), or 11 mutations (p.Glu148Gln, p.Pro369Ser, p.Phe479Leu,

Eur J Pediatr (2012) 171:1775–1785

p.Met680Ile, p.Ile692del, p.Met694Val, p.Met694Ile, p.Lys695Arg, p.Val726Ala, p.Ala744Ser, p.Arg761His) (n 01), or four mutations (p.Met694Val, p.Met680Ile, p.Met694Ile, p.Val726Ala) (n 01); (2) analysis of the whole gene in the case of heterozygosity. Six laboratories investigated in three steps: (1) exon 10 (n03), or exons 2 and 10 (n03); (2) exon 3 (n01), or exons 3 and 5 (n02), or exons 2 and 5 (n01), or exons 2, 3, and 5 (n02); and (3) sequencing of the remaining exons. Three laboratories always sequenced the entire gene not using a stepwise approach. Incidence calculation in Clinic-ESPED (n1) The number of cases reported simultaneously to both independent data sources (n1∩n2) and the remaining cases of n1 and n2 were used to determine the total number of definite FMF cases. The identification of cases reported to both ESPED surveys was possible by using the core data (see above) of the patients. The total number of cases was calculated using the following equation: n1Un2 ¼ ðn1 þ n2Þ ðn1 \ n2Þ . Only the number of cases reported in n1 was finally used as numerator to determine the minimum incidence of FMF with representation of a 95 % confidence interval, assuming a Poisson distribution. For calculation of the incidence in the entire study population, data of the Federal Office of Statistics were used as denominator (www.destatis.de). The peculiarity of this incidence calculation was the consideration of the migration background of the children. Because of the high proportion of Turkish children within our sample of FMF patients, we separately estimated the incidence in the population of children of Turkish origin ≤16 years of age in Germany. For this purpose, the Federal Office of Statistics provided data of the first-time micro-census surveys about children and adults with known migration background in Germany (www.destatis.de). The prevalence of FMF in Germany as well as for children of Turkish ancestry less than 16 years of age during the 3-year observation period was then calculated by multiplying the incidence per person-years with 16. Statistical analysis and data protection Frequency measurements were performed by descriptive analysis of each variable. Because of their seemingly nonGaussian distribution, continuous data are presented preferentially as medians and ranges. Discrete variables are described with proportional values. Personal data were pseudonymized and analyzed anonymously. With the information kept in the data bank, it was impossible to retrieve the identity of individuals. Data were also protected against unauthorized access. The study was

1777

approved by the ethics committee at the University of Düsseldorf. Parents and patients were instructed by an information letter.

Results Incidence of newly diagnosed FMF in Clinic-ESPED, n1 A total of 156 cases with HRF and among them, 122 (78 %) mutation-positive and symptomatic patients with FMF and 225 mutations were registered in n1 between July 2003 and June 2006. In the Laboratory-ESPED (n2), 544 symptomatic and asymptomatic cases as well as carriers with 804 mutations in the MEFV gene were documented. The core data revealed 73 duplicates with 134 mutations. Altogether, 593 different cases with 895 mutations were recorded in both surveillance systems (n1 U n2). We used the number of cases fulfilling the inclusion criteria for n1 for the minimum incidence calculation because the case definition in n2 could not be ensured in all children. The incidence of FMF per person-years in Germany was thus determined to be 3 per 106 person-years (95 % CI: 2.48–3.54 per 106 person-years) for the entire pediatric population and 55 per 106 person-years (95 % CI: 46–66 per 106 person-years) for children of Turkish ancestry in Germany (Table 1). Since FMF is a genetic disorder precluding recovery and unlikely to result in premature death before the age of 16 years, the prevalence of FMF in children under the age of 16 years may be calculated by multiplying the incidence per 106 person-years by 16, yielding an estimate of 48 (95 % CI: 39.7–55.6) per 106 children and 0.88 (95 % CI: 0.74–1.06) per 103 children of Turkish ancestry in Germany in this age group. Clinical presentation in Clinic-ESPED (n1) One hundred twenty-two children with FMF (68 male, 56 %; 54 female, 44 %) came from 120 families. Main countries of origin were Turkey (n092, 75 %), Lebanon (n08, 7 %), Armenia (n02, 2 %), and Iran (n02, 2 %). In seven (6 %) cases, the parents originated from different countries. In five (4 %) patients, the genetic background remained unknown (Fig. 1). A median age of onset of 4 (range 1–13) years and a delay of diagnosis of 2 years were calculated. Twenty-two (18 %) out of 122 children with symptomatic FMF were heterozygous for one mutation in the MEFV gene. The median duration of fever periods in this group was 3 (range 1–14) days. The main symptoms were abdominal pain (n014, 64 %), lymphadenopathy (n08, 36 %), arthralgia (n0 5, 23 %), and vomiting (n05, 23 %). Less frequent complaints were headache (n04, 18 %), arthritis, peritonitis, and pleuritis (each n03, 14 %), as well as chest pain, exanthema (each n0

1778

Eur J Pediatr (2012) 171:1775–1785

Table 1 Incidence of FMF in Clinic-ESPED (n1) FMF

July 2003–June 2004

July 2004–June 2005

July 2005–June 2006

Total number

All newly diagnosed symptomatic FMF cases in n1 Newly diagnosed symptomatic FMF cases of Turkish origin in n1 Population of children ≤16 years of age in Germany (2004–2006) Turkish population of children ≤16 years of age in Germany (2005–2007) Incidence per 1,000,000 person-years (all children in Germany) 95 % confidence interval (CI) Incidence per 1,000,000 person-years (Turkish children in Germany) 95 % CI

47 37 (79 %)

49 39 (80 %)

26 16 (62 %)

122 92 (75 %)

13 863 624

13 572 071

13 284 656

40 720 351a

587 200

554 200

520 600

1 662 000a

3.39 2.49–4.51 63.01

3.61 2.67–4.77 70.37

1.96 1.27–2.87 30.73

3 2.48–3.54 55

44.45–86.68

49.98–96.17

17.48–49.75

46–66

a

Cumulative person-years based on the annual number of children exposed; data of the Federal Office of Statistics

2, 9 %), myalgia, and diarrhea (each n01, 4.5 %). No case of amyloidosis was recorded. Colchicine was administered to 14 children (64 %). Four (18 %) and eight (36 %) patients received NSAIDs or no medicine at all. One hundred (82 %) out of 122 children were compound heterozygous (n050, 50 %), homozygous (n044, 44 %), or complex heterozygous (n06, 6 %). The median duration of fever periods was 2.7 (range 1–14) days. The main symptoms were abdominal pain (n077, 77 %), peritonitis, arthralgia (each n024, 24 %), arthritis (n020, 20 %), and vomiting (n016, 16 %). No case of amyloidosis was documented. Physical exercises were mentioned as a trigger factor (Fig. 2). Three children had no clinical manifestations. Therapeutic strategies differed considerably. Altogether, 83 (83 %) patients were treated with colchicine. Additionally, 28 (28 %) children received NSAIDs and 3 Fig. 1 Origin of children with FMF in the Clinic-ESPED survey (n10122)

(3 %) corticosteroids. Etanercept and methotrexate were added in individual cases. Six (6 %) and 17 (17 %) patients were treated only with NSAIDs or not at all. Genetics in Clinic—(n1) and Laboratory-ESPED (n2) Eighteen out of 23 laboratories participated regularly in this ESPED survey (see list in Acknowledgements). Senders were children’s hospitals and practices to 50 % each. Clinic-ESPED (n1) Two hundred twenty-five pyrin mutations were identified in the 122 patients. The most common mutations in the heterozygous group of 22 children were p.Met694Val (n012,

Eur J Pediatr (2012) 171:1775–1785

1779

Fig. 2 Clinical features of children with FMF in the Clinic-ESPED survey (n10100; homozygous n044, compound heterozygous n050, complex heterozygous n06; the heterozygous group n022 is not included in this figure)

54.5 %) and p.Glu148Gln (n07, 32 %). Other mutations were p.Val726Ala, p.Met680Ile, and p.Ala744Ser (each n0 1, 4.5 %). The 44 homozygous patients carried five different mutations: p.Met694Val (n030, 68 %), p.Met680Ile (n08, 18 %), p.Met694Ile (n04, 9 %), p.Glu148Gln (n03, 7 %), and p.Val726Ala (n01, 2 %). In the 50 compound heterozygotes, 13 different mutations were found (p.Met694Val, p.Met680Ile, p.Val726Ala, p.Arg761His, p.Met694Ile, p.Lys695Arg, p.Ala744Ser, p.Glu230Lys, p.Glu167Asp, p.Glu148Gln, p.Phe479Leu, p.Ile591Thr, and p.Thr267Ile). The most frequent mutations were p.Met694Val (n040, 40 %), p.Met680Ile (n022, 22 %), p.Val726Ala (n020, 20 %), p.Arg761His (n05, 5 %), and p.Met694Ile (n03, 3 %). The six most common complex heterozygous mutations were p.Pro369Ser/p.Arg408Gln (n 02, 14 %) and p.Leu110Pro/p.Glu148Gln (n 02, 14 %) followed by p.Met694Val/p.Arg761His plus p.Arg761His and p.Glu148Gln/p.Met694Val plus p.Glu148Gln.

Laboratory-ESPED (n2) Eight hundred four pyrin mutations were identified in 544 symptomatic and asymptomatic children with FMF as well as in asymptomatic carriers. Two hundred ninety-eight (37 %) heterozygous mutations, 240 (30 %) compound heterozygous mutations in 120 cases, 190 (24 %) homozygous mutations in 95 cases, and 76 (9 %) complex heterozygous mutations in 31 cases were recorded. n1 U n2 Seventy-three duplicates with 134 pyrin mutations were detected in both surveys. Altogether, 895 sequence variants in 23 different positions were identified in 593 children (Fig. 3). Three hundred seven (34 %) heterozygous cases, 284 (32 %) compound heterozygous mutations in 142 cases, 218 (24 %) homozygous mutations in 109 cases, and 86

1780

Eur J Pediatr (2012) 171:1775–1785

Fig. 3 The 23 pyrin sequence variants identified and their location with respect to the ten exons of the MEFV gene

(10 %) complex heterozygous mutations in 35 cases were recorded. The entire mutation spectrum is presented in Table 2 with p.Met694Val (n0402, 45 %) being the most common mutation, followed by p.Met680Ile (n 0129, 14 %), p.Val726Ala (n0107, 12 %), and p.Glu148Gln (n0103, 11.5 %).

Discussion To our knowledge, there are no published data on the incidence of FMF in Germany so far. Our prospective epidemiological Clinic-ESPED systematic survey found 156 newly diagnosed HRF patients, of whom 122 (78 %) had FMF. Based on this result, we estimated the minimum incidence of FMF in Germany as 3 per 106 person-years for the entire children population and as 55 per 106 personyears for children of Turkish descent living in Germany. The number of cases differs between different countries and between different regions within a country. The prevalence in Turkey was reported to be 0.1 % in the total population, 0.25 % in central Anatolia, and 0.82 % in Northern Turkey [17, 26, 28]. Mostly due to the Turkish immigrants, the disease has been calculated to affect around 4,000–8,000 individuals in Germany [16]. Our incidence calculation suggests about 25–50 newly diagnosed FMF cases per year. This translates into approximately 400–800 affected children ≤16 years, 300–600 of them being of Turkish origin. The risk thus is more than 15-fold higher in the population group with a Turkish background. Underreporting is a critical issue pertaining to our data as not all FMF patients might have been hospitalized or seen by a paediatric specialist. Additionally, our case definition excluded patients with FMF which had no defined mutation or lacked attacks of fever. Patients from private practices also have been included only in the Laboratory-ESPED (n2). Unfortunately, it was not clear for all cases of n2 whether the patients were asymptomatic, symptomatic, or only carriers. Thus, we could not estimate the incidence of n1 and n2 but only a minimum incidence of n1. Failure to report, failure to diagnose, and late diagnosis in adulthood

might also be issues contributing to the underestimation of the true incidence. The delay in correct diagnosis is due to the fact that symptoms manifested by patients with FMF are unspecific or uncommon or FMF as diagnosis is not considered. The incidence of FMF also varies in different geographic regions of Germany because of the different proportion of persons with a migration background, which is e.g., increased in Berlin and in the Rhein-Ruhr region. It has been postulated that analysis of the MEFV gene is of poor diagnostic value in western European Caucasians and that, therefore, a search for other HRFs should be initiated in affected patients [31]. Our results appear to confirm this supposition. In only three out of 122 cases, one parent was German or Swedish, but the other 119 children were of Mediterranean origin. Nevertheless, one has to be careful in interpreting these data, because German children with FMF may be severely underrepresented in this survey due to the fact that they do not have an “indicative” ancestry. This would be in line with the surprisingly high prevalence of the pyrin p.Glu148Gln and p.Lys695Arg substitutions of about 2 % each observed in the general German population which indicates that FMF may be underdiagnosed in German children with HRF [19]. FMF mutations might have conferred resistance to an as-yet-unknown pathogen as selective advantage, which would explain the high carrier frequency of mutations in some populations [1]. FMF is the only HRF with well-established diagnostic criteria for adults and children [18, 24, 38, 39]. In the first 2 years of life, the disease often begins with an atypical presentation, characterized by attacks of fever alone, and its diagnosis and initiation of treatment is therefore significantly delayed [29]. In Turkey and Iran, the mean ages of onset and diagnosis were 9.6±8.6 (11.4±7.8 years) and 16.4± 11.6 (12.4±11.8) years, respectively [6, 37]. In our survey, both parameters were substantially lower with a median of 4 and 6 years, respectively. In Turkish and Iranian children and adults, the cardinal signs of FMF were fever (90– 94.4 %), peritonitis (86.1–95 %), arthritis (27.7–47.4 %), pleuritis (31–47.2 %), myalgia (39.6 %), erysipelas-like erythema (5.6–20.9 %), and amyloidosis (5.6–12.9 %) [6, 37, 40]. In our Clinic-ESPED survey, in contrast, young

Eur J Pediatr (2012) 171:1775–1785 Table 2 Spectrum of 895 pyrin mutations in symptomatic and asymptomatic children in clinic—and laboratoryESPED (n1 U n2)

1781

All cases reported in n1 + n2 Heterozygous pyrin mutations p.Met694Val p.Glu148Gln p.Met680Ile p.Val726Ala p.Lys695Arg p.Arg761His p.Ala744Ser p.Met694Ile p.Ile591Thr p.Arg329His p.Glu230Lys p.Asn256Argfs*70 p.Ser141Ile Total number Homozygous pyrin mutations p.Met694Val p.Met680Ile p.Glu148Gln p.Met694Ile p.Val726Ala p.Glu230Lys p.Ala744Ser Total number

140 52 38 34 10 9 8 5 4 3 2 1 1 307

160 34 10 6 4 2 2 218

Compound heterozygous mutations, heterozygous complex and compound heterozygous complex alleles Compound heterozygous mutations p.Met694Val/p.Val726Ala 39 p.Met680Ile/p.Met694Val 37 p.Met680Ile/p.Val726Ala 9 p.Glu148Gln/p.Met694Val 9 p.Glu148Gln/p.Met680Ile 4 p.Met694Ile/p.Val726Ala 4 p.Met694Val/p.Arg761His 4 p.Thr267Ile/p.Val726Ala 4 p.Glu148Gln/p.Val726Ala 3 p.Met680Ile/p.Ala744Ser p.Ile591Thr/p.Met694Val p.Met694Val/p.Met694Ile p.Glu148Gln/p.Met694Ile p.Met694Val/p.Lys695Arg p.Met680Ile/p.Arg761His p.Val726Ala/p.Arg761His p.Phe479Leu/p.Val726Ala p.Glu167Asp/p.Val726Ala p.Glu148Gln/p.Glu148Val p.Met694Val/p.Glu230Lys p.Val726Ala/p.Ile591Thr p.Met694Val/p.Ala744Ser

3 3 3 2 2 2 2 2 2 1 1 1 1

1782

Eur J Pediatr (2012) 171:1775–1785

Table 2 (continued)

All cases reported in n1 + n2 p.Glu148Gln/p.Gly304Arg p.Glu148Gln/p.Ile591Thr p.Glu148Gln/p.Ala744Ser p.Val726Ala/p.Ser339Phe Total number Heterozygous complex alleles p.Pro369Ser/p.Arg408Gln p.Leu110Pro/p.Glu148Gln p.Glu148Gln/p.Pro369Ser/p.Arg408Gln p.Glu148Gln/p.Met694Ile p.Glu167Asp/p.Phe479Leu Total number Compound heterozygous complex alleles p.Glu148Gln/p.Pro369Ser/p.Arg408Gln + p.Ala457Val p.Glu148Gln/p.Val726Ala + p.Thr267Ile p.Glu148Gln/p.Met694Val + p.Met680Ile p.Met694Val/p.Arg761His + p.Arg761His p.Glu148Gln/p.Met694Ile + p.Met680Ile p.Glu148Gln/p.Met694Val + p.Glu148Gln Total number

children dominated and abdominal pain, lymphadenopathy, peritonitis, arthralgia, arthritis, and vomiting were the main clinical symptoms. Quite a few patients have clinical features differing from typical FMF, such as recurrent arthritis, recurrent myalgia, protracted febrile myalgia, bilateral erysipelas, and livedoid vasculopathy/atrophy blanche. Like FMF, such entities respond to colchicine treatment. The concomitant presence of pyrin mutations in other diseases also may modify their presentation and severity [4]. The dose and the position of the mutations in the MEFV gene have a major influence on the disease phenotype. One Turkish study demonstrated 17 different mutations in 32 % homozygous, 36 % compound heterozygous, and 28 % heterozygous patients. The mean symptom severity score and the Creactive protein levels were highest in the homozygous group [40]. In FMF patients of Arab descent, the highest disease severity was associated with p.Met694Val/p.Met694Val and p.Met694Val/p.Val724Ala, while p.Met694Ile/p.Met694Ile resulted in mild disease [15]. In another study, the p.Met694Val mutation also was associated with a more severe disease course characterized by earlier disease onset, more frequent attacks, higher prevalence of arthritis, pleuritis, and sacroiliitis, the need for higher doses of colchicines, and an increased risk of amyloidosis [13]. Carriage of the p.Met694Val, p.Met694Ile, and p.Met680Ile mutations in exon 10 thus usually predicts a severe clinical course, while a series of other mutations including p.Glu148Gln in exon 2 entails a mild course [23]. Our surveys present clear evidence that the p.Glu148Gln mutation is

1 1 1 1 284 14 6 2 2 3 56 3 2 1 1 1 1 30

nevertheless relevant for disease manifestation. In contrast, the clinical relevance of the p.Arg408Gln sequence variant, which is commonly observed in combination with p.Pro369Ser or p.Pro369Ser/p.Glu148Gln, remains unclear. Polymorphisms can be inherited in combination with a true FMF-associated mutation, either in trans or in cis (also known as a complex allele). Some studies have suggested a gene—dosage effect, with patients who have three or four MEFV mutations appearing to have more severe disease [1]. Environmental factors also appear to play an important role. Comparison of the disease activity of Turkish children living in Turkey with that of children of Turkish ancestry born and raised in Germany showed a significantly more severe disease course in patients residing in their native country [27]. In addition, ethnic differences are observed. For example, the p.Ala744Ser mutation is common in Arab populations, the p.Arg761His mutation is more frequent in Armenians, and the p.Phe479Leu mutation is commonly observed in Greeks [15]. The genotype—phenotype correlation thus is very complex, with ethnic and environmental factors playing a role in the clinical outcome. By the end of the year 2011, over 200 sequence variants (mutations and polymorphisms) in the MEFV gene have been collected in the INFEVERS data base (http:// fmf.igh.crns.fr/.ISSAID/invevers/) [25, 34]. Only five of these mutations (p.Met694Val, p.Val726Ala, p.Met680Ile, p.Met694Ile, and p.Glu148Gln) are responsible for about 70 % of the FMF cases [31]. For example, genetic analysis

Eur J Pediatr (2012) 171:1775–1785

of 1,090 Turkish patients revealed p.Met694Val as the most common mutation with 51 %, followed by p.Met680Ile (14 %) and p.Val726Ala (9 %) [37]. Although the studies are not comparable, our results are in agreement with these data, as the four most common mutations in Germany were p.Met694Val, p.Met680Ile, p.Val726Ala, and p.Glu148Gln. Amyloidosis is still a substantial health problem for adult FMF patients. The “International Study group for Phenotype—Genotype correlation in FMF” analyzed 2,482 cases, among which were 260 patients with amyloidosis. The key risk factor for the development of amyloidosis was the country of recruitment, followed by p.Met694Val homozygosity [35]. After the introduction of treatment with colchicine, a decrease in the percentage of biopsies with secondary amyloidosis from 12.1 % (1978–1990) to 2 % (2000–2009) was observed. The main reason for this was an increased awareness of the disease and a better medical care of affected patients [2]. Decision trees and scores may help in the diagnostic evaluation of children with HRF. These diagnostic tools may optimize molecular analyses by suggesting the order in which genes should be screened as well as improve the assessment of disease activity in HRF [10, 11, 27, 30]. As a step forward in this direction, a German clinical and research consortium (AID-Net) was established in 2009, including an online-registry for AID (accessible via http:// www.aid-register.uk-essen.de), biomaterial banks (DNA/serum), and basic research projects that focus on molecular mechanisms of AID [20]. The translational approach of AID-Net combines projects regarding epidemiology, clinical and immunological features, as well as molecular genetics.

Conclusions This first systematic approach to estimate the incidence of FMF in Germany yielded a minimum incidence of 3 per 106 person-years for the entire population of children and of 55 per 106 person-years for children of Turkish descent. The results correspond to about 25–50 newly diagnosed young patients per year. FMF is most common among immigrants, and its frequency thus depends on the geographic region in Germany. In children with the suspicion of FMF, an accurate clinical history including ethnic ancestry, and a detailed family history as well as a physical examination are the initial diagnostic steps, before genetic analysis of the primarily affected exons 2, 3, 9, and 10 is performed, while a sequence analysis of the whole MEFV gene is only necessary in heterozygous cases. The p.Met694Val, p.Met680Ile, p.Val726Ala, and p.Glu148Gln mutations are the prevailing pyrin mutations. Importantly, 18 % of the clinically manifested FMF cases carried the disease-causing mutation only

1783

in heterozygous form, consistent with the current understanding that FMF is an autosomal codominantly inherited disease [9]. Acknowledgments The authors thank the following colleagues for their collaboration: Clinic-ESPED n0122: T. Kallinich (Charité, Berlin) n022 I. Földvari (Clinic Eilbeck, Hamburg) n011 U. Kruse (University children’s hospital, Essen) n010 P. Lankisch (HHU Children’s hospital, Düsseldorf) n05 K. Mönkemöller (Children’s hospital, Köln) n05 C. Rietschel (Clementine-Children’s hospital, Frankfurt) n05 S. Buderus, C. Franz (St. Marien-Hospital, Bonn) n04 H. Wittkowski, D. Föll (University children’s hospital, Münster) n04 T. Berger (Children’s hospital, Datteln) n03 M. Gabriel, B. Markert (Clinical centre, Aschaffenburg) n03 E. Lankes, A. Pfünder (Children’s hospital Schwabing, München) n03 H. Bachmann, J. Dallmeyer (Hospital Links der Weser, Bremen) n02 U. Brand (Clinical centre, Ludwigsburg) n02 S. De Negri (Hospital Dritter Orden, München) n02 T. Hospach (Olgahospital, Stuttgart) n02 J. Kümmerle-Deschner (University children’s hospital, Tübingen) n02 K. Minden (HELIOS hospital, Berlin-Buch) n02 R. Rossi, J. Rakob (Hospital Neukölln, Berlin) n02 S. Wangemann (St. Johannes-Hospital, Duisburg) n02 U. Winckelmann, GM. Beron (Dr.-H.-Schmidt-Hospital, Wiesbaden) n02 S. Wirth, (HELIOS hospital, Wuppertal) n02 F. Zepp (University children’s hospital, Mainz) n02 N. Baumgärtner (University children’s hospital, Frankfurt) n01 S. Benseler (University children’s hospital, Bonn) n01 H. Böhme (Harz-Clinical centre GmbH, Wernigerode) n01 P. Conjeevaram (Clinical centre, Offenbach) n01 M. Erdmann (Marienkrankenhaus, Papenburg) n01 U. Ermer (Hospital St. Elisabeth, Neuburg/Donau) n01 S. Fahl (University children’s hospital, Köln) n01 G. Ganser (St. Josef-Stift, Sendenhorst) n01 K. Huß (LMU Children’s hospital, München) n01 B. Kinder (Clinical centre, Neubrandenburg) n01 H. Koch (St. Marien-Hospital, Vechta) n01 A. Längler (General hospital, Herdecke) n01 J. Möller (Clinical centre, Saarbrücken) n01 C. Müller (General hospital, Solingen) n01 W. Rauh (Mother house of Borromeans, Trier) n01 J. Rösler (University children’s hospital, Dresden) n01 M. Scharnetzky (Diakonie Hospital, Rotenburg) n01 U. Schimmel (Children’s hospital, Hagen) n01 P. Schlumberger (University children’s hospital, Marburg) n01 L. Schmid (County hospital, Altötting) n01 M. Schulze Becking (Clinical centre, Oldenburg) n01 J. Spiegler (University children’s hospital Schleswig-Holstein, Lübeck) n01 V. Stephan (St. Josef-Hospital, Bochum) n01 F.K. Trefz (County Hospital, Reutlingen) n01 C. Willaschek (Caritas-Hospital, Bad Mergentheim) n01 Laboratory-ESPED n0544: S. Burwinkel (Department of Molecular Medicine, Bernhard Nocht Institute for Tropical Medicine, Hamburg) n0172 H. Ruebsamen (Department of Clinical Chemistry – Großhadern, University of Munich, Munich) n0108 T. Haverkamp (MVZ Dortmund Dr. Eberhard & Partner, Dortmund) n096 G. Wildhardt (Bioscientia Center for Human Genetics, Ingelheim) n040 T. Trosch, D. Gläser (Genetics, Dr. Mehnert & Partner, Neu-Ulm) n035 C. Aulehla-Scholz (Institute of Clinical Genetics, Olgahospital, Stuttgart) n025 S. Kleinle (synlab MVZ Human Genetics München GmbH, München) n023 R. Maiwald (MVZ for Laboratory Medicine, Mönchengladbach) n018

1784 F. Austrup (Laboratory for Human Genetics, Datteln) n07 E. Krasemann (MVZ Lab Fenner & Colleagues, Hamburg) n07 C. Meyer-Kleine (Center for Human Genetics, Linden and laboratory Beudt, Frankfurt) n04 T. Rogge (Laboratory of the Vivantes-Klinikum Neukölln, Berlin) n03 H. Skladny (Center for Human Genetics, Mannheim) n02 S. Theil (Institute for Tropical Medicine, University Tübingen, Tübingen) n02 H. Gabriel (Center for Human Genetics, Osnabrück) n01 W. Schmidt (Laboratory Arndt, Hamburg) n01 No cases of W. Heinritz (Institute for applied human genetics and oncogenetics, Leipzig) The authors thank U. Göbel and B. Heinrich from the ESPED registry. E. Lainka gratefully acknowledges the comments made by O. Weiergräber (Forschungszentrum Jülich, Germany). Conflict of interest E. Lainka and all coauthors declare that there are no conflicts of interest, relationships, and affiliations relevant to this manuscript.

References 1. Aksentijevich I, Kastner DL (2011) Genetics of monogenic autoinflammatory diseases: past successes, future challenges. Nat Rev Rheumatol 7(8):469–478 2. Akse-Onal V, Sag E, Ozen S et al (2010) Decrease in the rate of secondary amyloidosis in Turkish children with FMF: are we doing better? Eur J Pediatr 169:971–974 3. Ben-Chetrit E, Levy M (1998) Familial Mediterranean fever. Lancet 351:659–664 4. Ben-Chetrit E, Peleg H, Aamar S et al (2009) The spectrum of MEFV clinical presentations—is it familial Mediterranean fever only? Rheumatology 48:1455–1459 5. Ben-Chetrit E, Touitou I (2009) Familial Mediterranean Fever in the world. Arthritis Rheum 61(10):1447–1453 6. Bidari A, Ghavidel-Parsa B, Najmabadi H et al (2010) Common MEFV mutation analysis in 36 Iranian patients with familial Mediterranean fever: clinical and demographic significance. Mod Rheumatol 20:566–572 7. Booth DR, Gillmore JD, Lachmann HJ et al (2000) The genetic basis of autosomal dominant familial Mediterranean fever. QJM 93:217–221 8. Chae JJ, Aksentijevich I, Kastner DL (2009) Advances in the understanding of familial Mediterranean fever and possibilities for targeted therapy. Br J Haematol 146:467–478 9. Chae JJ, Cho YH, Lee GS, Cheng J, Liu PP, Feigenbaum L, Katz SI, Kastner DL (2011) Gain-of-function pyrin mutations induce NLRP3 protein-independent interleukin-1ß activation and severe autoinflammation in mice. Immunity 34:755–768 10. Federici L, Rittore-Domingo C, Kone-Paut I et al (2006) A decision tree for genetic diagnosis of HPF in unselected patients. Ann Rheum Dis 65:1427–32 11. Gattorno M, Sormani MP, D’Osualdo A et al (2008) A diagnostic score for molecular analysis of hereditary autoinflammatory syndromes with periodic fever in children. Arthritis Rheum 58:1823– 1832 12. Goldfinger SE (1972) Colchicine for familial Mediterrean fever. N Engl J Med 287:1302 13. Grateau G (2000) The relation between familial Mediterranean fever and amyloidosis. Curr Opin Rheumatol 12:61–64 14. Gross O, Thomas CJ, Guarda G et al (2011) The inflammasome. An integrated view. Immunol Rev 243(1):136–151 15. Guz G, Kanbay M, Ozturk MK (2009) Current perspectives on familial Mediterranean fever. Curr Opin Infect Dis 22:309– 315

Eur J Pediatr (2012) 171:1775–1785 16. Kallinich T, Keitzer R (2007) Periodic fever syndromes. In: Wagner N, Dannecker G (eds) Paediatric Rheumatology. Springer medicine publishing, Heidelberg, pp 414–421 17. Kisacik B, Yildirim B, Tasliyurt T (2009) Increased frequency of familial Mediterranean fever in northern Turkey: a population based study. Rheumatol Int 11:1307–1309 18. Kondi A, Hentgen V, Piram M et al (2010) Validation of the new paediatric criteria for the diagnosis of familial Mediterranean fever: data from a mixed population of 100 children from the French reference centre for auto-inflammatory disorders. Rheumatology 49:2200–2203 19. Kümpfel T, Gerdes LA, Wacker T et al. (2012) Familial Mediterranean fever-associated mutation pyrin E148Q as a potential risk factor for multiple sclerosis. Mult Scler (in press) 20. Lainka E, Bielak M, Hilger V et al (2011) Translational research network and patient registry for autoinflammatory disease. Rheumatology 50:237–242 21. Lainka E, Neudorf U, Lohse P et al (2010) Analysis of cryopyrinassociated periodic syndromes (CAPS) in German children: epidemiological, clinical and genetic characteristics. Klin Padiatr 222:356–361 22. Lidar M, Scherrmann JM, Shinar Y et al (2004) Cochicine nonresponsiveness in familial Mediterranean fever: clinical, genetic, pharmacokinetic, and socioeconomic characterization. Semin Arthritis Rheum 33:273–282 23. Livneh A (2011) Familial Mediterranean fever: a continuously challenging disease. IMAJ 13:197–198 24. Livneh A, Langeviz P, Zemer D (1997) Criteria for the diagnosis of familial Mediterranean fever. Arthritis Rheum 40:1879–1885 25. Milhavet F, Cuisset L, Hoffman HM et al (2008) The infevers autoinflammatory mutation online registry: update with new genes and functions. Hum Mutat 29:803–808 26. Onen F, Sumer H, Turkay S et al (2004) Increased frequency of familial Mediterranean fever in Central Anatolia, Turkey. Clin Exp Rheumatol 22(4 Suppl 34):S31–3 27. Ozen S, Aktay N, Lainka E et al (2009) Disease severity in children and adolescents with familial Mediterranean fever: a comparative study to explore environmental effects on a monogenic disease. Ann Rheum Dis 68:246–248 28. Ozen S, Frenkel J, Ruperto N et al (2011) The Eurofever project: towards better care for autoinflammatory diseases. Eur J Pediatr 170:445–452 29. Padeh S, Livneh A, Pras E et al (2010) Familial Mediterranean fever in the first 2 years of life: a unique phenotype of disease in evolution. J Pediatr 156:985–989 30. Piram M, Frenkel J, Gattorno M et al (2011) A preliminary score for the assessment of disease activity in hereditary recurrent fevers: results from the AIDAI (Auto-Inflammatory Diseases Activity Index) Consensus Conference. Ann Rheum Dis 70:309–314 31. Tchernitchko D, Moutereau S, Delahaye A et al (2005) MEFV analysis is of particularly weak diagnostic value for recurrent fevers in western European Caucasian patients. Arthritis Rheum 52:3603–3605 32. The French FMF Consortium (1997) A candidate gene for familial Mediterranean fever. Nat Genet 17:25–31 33. The International FMF Consortium (1997) Ancient missense mutations in a new member of the RoRet gene family are likely to cause familial Mediterranean fever. Cell 90:797–807 34. Touitou I, Lesage S, Mc Dermott M et al (2004) Infevers: an evolving mutation database for autoinflammatory syndromes. Human Mutat 24:194–198 35. Touitou I, Sarkisian T, Medlej-Hashim M et al (2007) Country as the primary factor for renal amyloidosis in familial Mediterranean fever. Arthritis Rheum 56:1706–1712

Eur J Pediatr (2012) 171:1775–1785 36. Tsuchiya-Suzuki A, Yazaki M, Nakamura A et al (2009) Clinical and genetic features of familial Mediterranean fever in Japan. J Rheumatol 36(8):1671–1676 37. Tunca M, Akar S, Onen F (2005) Familial Mediterrean fever (FMF) in Turkey: results of a nationwide multicenter study. Medicine (Baltimore) 84:1–11 38. Ureten K, Göndülalan S, Akbal E et al (2010) Demographic, clinical and mutational characteristics of Turkish familial Mediterranean fever patients: results of a single center in Central Anatolia. Rheumatol Int 30:911–915

1785 39. Yalcinkaya F, Özen S, Özcakar B et al (2009) A new set of criteria for the diagnosis of familial Mediterranean fever in childhood. Rheumatology 48:395–8 40. Yilmaz R, Ozer S, Ozyurt H et al (2009) Familial Mediterranean fever gene mutations in the inner northern region of Turkey and genotype-phenotype correlation in children. J Paediatr Child Health 45:641–645 41. Zemer D, Pras M, Sohar E et al (1986) Colchicine in the prevention and treatment of the amyloidosis of familial Mediterranean fever. N Engl J Med 314:1001–1005