familial mediterranean fever among patients from karabakh and the diagnostic value of mefv gene...

8
ARTHRITIS & RHEUMATISM Vol. 48, No. 8, August 2003, pp 2324–2331 DOI 10.1002/art.11102 © 2003, American College of Rheumatology Familial Mediterranean Fever Among Patients From Karabakh and the Diagnostic Value of MEFV Gene Analysis in All Classically Affected Populations Ce ´cile Cazeneuve, 1 Zaruhi Hovannesyan, 2 David Genevie `ve, 3 Hasmik Hayrapetyan, 2 Ste ´phanie Papin, 4 Emmanuelle Girodon-Boulandet, 3 Brigitte Boissier, 3 Josue ´ Feingold, 4 Karine Atayan, 5 Tamara Sarkisian, 2 and Serge Amselem 1 Objective. Familial Mediterranean fever (FMF) is an autosomal-recessive disorder that is common in Armenian, Turkish, Arab, and Sephardic Jewish popu- lations. Its clinical diagnosis is one of exclusion, with the patients displaying nonspecific symptoms related to serosal inflammation. MEFV gene analysis has provided the first objective diagnostic criterion for FMF. How- ever, in the absence of an identified mutation (NI/NI genotype), both the sensitivity of the molecular analyses and the involvement of the MEFV gene in FMF are called into question. The present study was designed to further evaluate the diagnostic value of MEFV analysis in another population of Mediterranean extraction. Methods. The MEFV gene was screened for mu- tations in 50 patients living in Karabakh (near Arme- nia) who fulfilled the established criteria for FMF. In addition, we analyzed published series of patients from the above-mentioned at-risk populations. Results. The mutation spectrum in Karabakhian patients, which consisted of only 6 mutations (with 26% of NI alleles), differed from that reported in Armenian patients. Strikingly, among patients from Karabakh and among all classically affected populations, the distribution of genotypes differed dramatically from Hardy-Weinberg equilibrium (P 0.0016 and P < 0.00001, respectively). These results, combined with other population genetics–based data, revealed the ex- istence of an FMF-like condition that, depending on the patients’ ancestry, was shown to affect 85–99% of those with the NI/NI genotype. Conclusion. These data illuminate the meaning of negative results of MEFV analyses and show that in all populations evaluated, most patients with the NI/NI genotype had disease that mimicked FMF and was unrelated to the MEFV gene. Our findings also demon- strate the high sensitivity of a search for very few mutations in order to perform a molecular diagnosis of MEFV-related FMF. One major aim of the first molecular studies performed in populations affected by familial Mediter- ranean fever (FMF) was to test the diagnostic value of the analysis of MEFV (1,2), a strong candidate gene for this autosomal-recessive condition (MIM 249100) that primarily affects Armenian, Sephardic Jewish, Turkish, and Arab populations (3). The diagnosis of FMF is indeed one of exclusion, based on established clinical criteria (4). FMF patients display recurrent episodes of fever and serosal inflammation (manifested by nonspe- cific signs, such as sterile arthritis, peritonitis, and/or pleurisy) associated with an increased erythrocyte sedi- mentation rate and increased serum levels of acute- phase proteins. However, establishing the diagnosis of Supported by grants from NATO (grant LST.CLG.978883), AFM/INSERM (Re ´seau de Recherche sur les Maladies Rares), and from the Calouste Gulbenkian Foundation (Lisbon, Portugal). Dr. Cazeneuve is the recipient of a fellowship from INSERM. 1 Ce ´cile Cazeneuve, PharmD, PhD, Serge Amselem, MD, PhD: Service de Biochimie et de Ge ´ne ´tique Mole ´culaire and INSERM (Unite ´ 468), Ho ˆpital Henri-Mondor, Cre ´teil, France; 2 Zaruhi Hovan- nesyan, PhD, Hasmik Hayrapetyan, PhD, Tamara Sarkisian, PhD: Center of Medical Genetics, National Academy of Sciences, Yerevan, Armenia; 3 David Genevie `ve, MD, Emmanuelle Girodon-Boulandet, MD, Brigitte Boissier, BSc: Ho ˆpital Henri-Mondor, Cre ´teil, France; 4 Ste ´phanie Papin, MS, Josue ´ Feingold, MD: INSERM (Unite ´ 468), Ho ˆpital Henri-Mondor, Cre ´teil, France; 5 Karine Atayan, MD: Step- anakert Clinic, Stepanakert, Karabakh. Address correspondence and reprint requests to Serge Am- selem, MD, PhD, Service de Biochimie et de Ge ´ne ´tique Mole ´culaire, INSERM Unite ´ 468, Ho ˆpital Henri-Mondor, 51 Avenue du Mare ´chal de-Lattre-de-Tassigny, Cre ´teil F-94010, France. E-mail: serge.amselem @im3.inserm.fr. Submitted for publication January 30, 2003; accepted in revised form April 14, 2003. 2324

Upload: cecile-cazeneuve

Post on 06-Jun-2016

214 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Familial Mediterranean fever among patients from Karabakh and the diagnostic value of MEFV gene analysis in all classically affected populations

ARTHRITIS & RHEUMATISMVol. 48, No. 8, August 2003, pp 2324–2331DOI 10.1002/art.11102© 2003, American College of Rheumatology

Familial Mediterranean Fever Among Patients FromKarabakh and the Diagnostic Value of

MEFV Gene Analysis in All Classically Affected Populations

Cecile Cazeneuve,1 Zaruhi Hovannesyan,2 David Genevieve,3 Hasmik Hayrapetyan,2

Stephanie Papin,4 Emmanuelle Girodon-Boulandet,3 Brigitte Boissier,3 Josue Feingold,4

Karine Atayan,5 Tamara Sarkisian,2 and Serge Amselem1

Objective. Familial Mediterranean fever (FMF) isan autosomal-recessive disorder that is common inArmenian, Turkish, Arab, and Sephardic Jewish popu-lations. Its clinical diagnosis is one of exclusion, withthe patients displaying nonspecific symptoms related toserosal inflammation. MEFV gene analysis has providedthe first objective diagnostic criterion for FMF. How-ever, in the absence of an identified mutation (NI/NIgenotype), both the sensitivity of the molecular analysesand the involvement of the MEFV gene in FMF arecalled into question. The present study was designed tofurther evaluate the diagnostic value of MEFV analysisin another population of Mediterranean extraction.

Methods. The MEFV gene was screened for mu-tations in 50 patients living in Karabakh (near Arme-nia) who fulfilled the established criteria for FMF. Inaddition, we analyzed published series of patients fromthe above-mentioned at-risk populations.

Results. The mutation spectrum in Karabakhianpatients, which consisted of only 6 mutations (with 26%of NI alleles), differed from that reported in Armenianpatients. Strikingly, among patients from Karabakhand among all classically affected populations, thedistribution of genotypes differed dramatically fromHardy-Weinberg equilibrium (P � 0.0016 and P <0.00001, respectively). These results, combined withother population genetics–based data, revealed the ex-istence of an FMF-like condition that, depending on thepatients’ ancestry, was shown to affect 85–99% of thosewith the NI/NI genotype.

Conclusion. These data illuminate the meaning ofnegative results of MEFV analyses and show that in allpopulations evaluated, most patients with the NI/NIgenotype had disease that mimicked FMF and wasunrelated to the MEFV gene. Our findings also demon-strate the high sensitivity of a search for very fewmutations in order to perform a molecular diagnosis ofMEFV-related FMF.

One major aim of the first molecular studiesperformed in populations affected by familial Mediter-ranean fever (FMF) was to test the diagnostic value ofthe analysis of MEFV (1,2), a strong candidate gene forthis autosomal-recessive condition (MIM 249100) thatprimarily affects Armenian, Sephardic Jewish, Turkish,and Arab populations (3). The diagnosis of FMF isindeed one of exclusion, based on established clinicalcriteria (4). FMF patients display recurrent episodes offever and serosal inflammation (manifested by nonspe-cific signs, such as sterile arthritis, peritonitis, and/orpleurisy) associated with an increased erythrocyte sedi-mentation rate and increased serum levels of acute-phase proteins. However, establishing the diagnosis of

Supported by grants from NATO (grant LST.CLG.978883),AFM/INSERM (Reseau de Recherche sur les Maladies Rares), andfrom the Calouste Gulbenkian Foundation (Lisbon, Portugal). Dr.Cazeneuve is the recipient of a fellowship from INSERM.

1Cecile Cazeneuve, PharmD, PhD, Serge Amselem, MD,PhD: Service de Biochimie et de Genetique Moleculaire and INSERM(Unite 468), Hopital Henri-Mondor, Creteil, France; 2Zaruhi Hovan-nesyan, PhD, Hasmik Hayrapetyan, PhD, Tamara Sarkisian, PhD:Center of Medical Genetics, National Academy of Sciences, Yerevan,Armenia; 3David Genevieve, MD, Emmanuelle Girodon-Boulandet,MD, Brigitte Boissier, BSc: Hopital Henri-Mondor, Creteil, France;4Stephanie Papin, MS, Josue Feingold, MD: INSERM (Unite 468),Hopital Henri-Mondor, Creteil, France; 5Karine Atayan, MD: Step-anakert Clinic, Stepanakert, Karabakh.

Address correspondence and reprint requests to Serge Am-selem, MD, PhD, Service de Biochimie et de Genetique Moleculaire,INSERM Unite 468, Hopital Henri-Mondor, 51 Avenue du Marechalde-Lattre-de-Tassigny, Creteil F-94010, France. E-mail: [email protected].

Submitted for publication January 30, 2003; accepted inrevised form April 14, 2003.

2324

Page 2: Familial Mediterranean fever among patients from Karabakh and the diagnostic value of MEFV gene analysis in all classically affected populations

FMF is essential, mainly because the signs and symp-toms of FMF may lead to unnecessary exploratorysurgery. Moreover, daily administration of colchicinecontrols the symptoms in most patients with FMF (5,6)and, most important, prevents the occurrence of renalamyloidosis (7), the main complication of FMF, whoseoccurrence is influenced by MEFV genotype and MEFV-independent modifying genetic factors (8).

In the course of the first molecular investigationsperformed in patients with FMF, we showed that char-acterization of mutated MEFV alleles represents the firstobjective diagnostic criterion for this disease. Indeed, weidentified 2 MEFV allele mutations in 89% of a largeand unselected cohort of unrelated Armenian patients(9), who fulfilled the established clinical criteria for thediagnosis of FMF. It was noteworthy, however, that in11% of the patients, either no mutation was detected oronly 1 FMF allele was characterized, even after thescreening of all 10 MEFV coding regions and intronboundaries for mutations by denaturing gradient gelelectrophoresis (DGGE), a highly sensitive techniquefor the detection of mutations (10). This latter observa-tion left open the possibility that, in some cases, the signsand symptoms of FMF documented in patients of Ar-menian ancestry did not result from mutations in theMEFV gene. Thus, an investigation of other affectedpopulations might provide information of critical impor-tance in confirming or excluding this hypothesis. Exceptfor our previous study (9), however, none of the studiesperformed thus far in large homogeneous populationswas based on exhaustive screening for mutations of allMEFV exons. Indeed, the latter studies focused on thesearch for few known mutations, and, since all of thestudies described patients with uncharacterized FMFalleles, the possibility that MEFV mutations remainedundetected cannot be ruled out.

In the present study, we further tested the diag-nostic value of MEFV gene analysis and delineated themutation spectrum in another population of FMF pa-tients of Mediterranean extraction. We evaluated acohort of unrelated patients living in Karabakh, which isnear Armenia and is inhabited mainly by people ofArmenian ancestry. As in Armenia, the prevalence ofFMF in this area is high (Atayan K, Sarkisian T:unpublished observations). Considering the frequency ofuncharacterized MEFV alleles, the unexpectedly lownumber of Karabakhian patients with only 1 identifiedMEFV mutation led us to test whether the MEFVgenotype distribution was consistent with Hardy-Weinberg equilibrium (i.e., binomial distribution). Thedata obtained prompted us to perform a detailed analy-

sis of 4 previous reports on various populations classi-cally affected by FMF: subjects of Armenian (9), Sephar-dic Jewish (11), Turkish (12), and Arab (13) extraction.The results of this evaluation challenge both the diag-nostic value of MEFV gene analysis in all tested popu-lations and the molecular basis of this hereditary inflam-matory disorder.

PATIENTS AND METHODS

Patients. We evaluated 50 consecutive FMF patientsfrom 39 unrelated families living in Karabakh, representing atotal of 84 independent alleles. All of the patients belong tofamilies in which there was no known consanguinity. None ofthe families was selected through genetic linkage analysesusing MEFV gene markers. The age range of the patients was4–71 years (mean 25.0 years). The ratio of males to femaleswas 0.61. Clinical features were recorded by the same investi-gator (KA) using a standardized form. The diagnosis of FMFwas made according to established clinical criteria (4). In-formed consent was given by all patients or their parents.

Mutation analysis. The mutation analysis was per-formed as described previously (9). Briefly, the M694V andV726A mutations were searched for by restriction enzymeanalyses. MEFV gene analysis was completed in patients whohad, at most, only 1 of these 2 mutations, by DGGE screeningof all 10 coding exons and flanking intron sequences. Samplesdisplaying a shift in mobility were directly sequenced.

Statistical analysis. The spectrum of MEFV genemutations in Karabakh and in Armenia (9) was compared withthe chi-square test (using the Yates’ adjustment when appro-priate). We calculated p1 and q1, the frequency of identified (I)and nonidentified (NI) mutations among Karabakhian pa-tients, using the classic counting method.

p1 ��2nI/I � nI/NI�

2NTotal

and

q1 � 1 � p1 ��2nNI/NI � nI/NI�

2NTotal

where nI/I is the number of patients carrying an I/I genotype,nI/NI the number carrying an I/NI genotype, nNI/NI the numbercarrying an NI/NI genotype, and NTotal is the total number ofpatients (NTotal � nI/I � nI/NI � nNI/NI).

The expected distributions of I/I, I/NI, and NI/NIgenotypes were then calculated according to the assumption ofHardy-Weinberg equilibrium. The observed and expected dis-tributions of the 3 genotypes were compared using the chi-square test. Similar calculations were performed from the datain 4 published reports that focused on the investigation ofpatients belonging to populations classically affected by FMF(i.e., Armenians [9], Sephardic Jews [11], Turks [12], andArabs [13]).

The results of these statistical analyses, combined withpopulation genetics–based data (see Discussion), led us toconclude that the signs and symptoms of FMF presented byseveral patients did not result from mutations in the MEFV

FMF IN PATIENTS OF KARABAKH AND MEFV GENE ANALYSIS 2325

Page 3: Familial Mediterranean fever among patients from Karabakh and the diagnostic value of MEFV gene analysis in all classically affected populations

gene. To evaluate the number of these latter patients (NOther),we first calculated the number of patients whose phenotype didresult from mutations in MEFV (NMEFV). To this end, weconsidered that the symptoms were related to MEFV inpatients carrying 2 mutated alleles (I/I genotype). In addition,assuming that some mutated alleles might have escaped ourmolecular screening procedure, we postulated that the symp-toms were also related to the MEFV gene in all patientscarrying an I/NI genotype and in an undetermined number ofpatients carrying an NI/NI genotype; the NI alleles in thesepatients therefore correspond to uncharacterized mutatedMEFV alleles. Designating p2 as the frequency of identifiedmutations and q2 as the frequency of nonidentified mutationamong the NMEFV patients, with q2 � 1 – p2 and p1 � p2, then

1.

nI/I � p22NMEFV NMEFV �

nI/I

p22

and

2.

nI/NI � 2p2q2NMEFV � 2p2 (1 � p2)NMEFV

NMEFV �nI/NI

2p2 (1 � p2)

Given 1 and 2, we have the following:

nI/I

p22 �

nI/NI

2p2 �1 � p2�

2p2 �1 � p2�nI/I � p22nI/NI

p2nI/NI � 2nI/I�1 � p2� � 0

p2 �2nI/I � nI/NI� � 2nI/I

p2 �2nI/I

2nI/I � nI/NI�

nI/I

nI/I �nI/NI

2

Thus, NMEFV is equal to

NMEFV �nI/I

p22 �

nI/I

�nI/I

nI/I �nI/NI

2 �2 �

�nI/I �nI/NI

2 �2

nI/I

and NOther is equal to

NOther � NTotal � NMEFV

RESULTS

The spectrum of MEFV gene mutations in pa-tients from Karabakh is detailed in Table 1. Six differentmutations and several polymorphisms (data not shown)were identified in our population sample. These muta-tions accounted for 74% of the 84 independent FMFalleles. M694V, M680I, and V726A substitutions repre-sented 68%, 16%, and 5% of the 62 characterized

alleles, respectively. Thirteen different genotypes werecharacterized among 45 of the 50 patients (Table 2).Both FMF alleles were characterized in 30 of the 45patients (67%). Both FMF alleles remained uncharac-terized in 7 patients, whereas 1 FMF allele was unchar-acterized in 8 patients (Table 2).

We then compared the spectrum of MEFV genemutations among Karabakhian patients with the spec-trum reported among Armenian patients (9). The fre-quency of nonidentified MEFV alleles was much higherin patients from Karabakh than in patients from Arme-nia (26% [22 of 84] versus 7% [12 of 163]; P � 0.00005)(Figure 1A). The spectrum of identified mutations inKarabakhian and Armenian patients also differed (Fig-ure 1B). Indeed, the M694V and R761H mutations weremore frequent among Karabakhian patients (P � 0.01and P � 0.04, respectively). Conversely, the V726Amutation represented only 5% of the identified alleles in

Table 1. Spectrum of MEFV gene mutations in patients fromKarabakh

MutationNo. (%) of

independent alleles

M694V 42 (50.0)M680I* 10 (11.9)R761H 4 (4.8)V726A 3 (3.6)E148Q 2 (2.4)M694I 1 (1.2)NI† 22 (26.2)Total 84 (100)

* Nucleotide 2040, ATG�ATC.† Nonidentified (NI) allele.

Table 2. Genotypes at the MEFV locus in patients from Karabakh

Genotype*No. (%)

of patients

M694V/M694V 15 (33)M694V/M680I 7 (16)M694V/R761H 2 (4)M694V/V726A 2 (4)M680I/M680I 1 (2)M680I/R761H 1 (2)V726A/R761H 1 (2)M694V/E148Q 1 (2)M694V/NI 4 (9)M680I/NI 2 (4)M694I/NI 1 (2)E148Q/NI 1 (2)NI/NI 7 (16)Total 45 (100)†

* NI � nonidentified allele.† These 45 patients included 39 independent patients and 6 patientswho shared 1 allele with an affected relative.

2326 CAZENEUVE ET AL

Page 4: Familial Mediterranean fever among patients from Karabakh and the diagnostic value of MEFV gene analysis in all classically affected populations

Karabakhian patients, whereas it accounted for 26% ofthe characterized FMF alleles in Armenian patients(P � 0.0003). Only 1 mutation (M680I) was found at asimilar frequency in the 2 populations. No new MEFVmutation was identified among the Karabakhian pa-tients.

One striking observation was the relatively lownumber of Karabakhian patients with only 1 identifiedMEFV mutation and the relatively high number ofpatients with 2 uncharacterized MEFV alleles (Table 2),with respect to the high frequency of uncharacterizedFMF alleles (Table 1). This unexpected findingprompted us to test whether the distribution of I/I (2identified MEFV mutations), I/NI (1 identified MEFVmutation), and NI/NI (no identified MEFV mutation)genotypes complied with Hardy-Weinberg equilibrium.

This analysis revealed that the distribution ofgenotypes among Karabakhian patients differed signifi-cantly from a binomial distribution (P � 0.0016) (Figure2). A similar analysis performed in the neighboringArmenian population of patients demonstrated thesame departure from Hardy-Weinberg equilibrium (P �0.00001) (Figure 2), whereas the distribution of the mostcommon MEFV mutations in these 2 populations did notsignificantly differ from Hardy-Weinberg expectations(P � 0.053 and P � 0.796 for M694V and M680I inKarabakhian patients; P � 0.092, P � 0.446, and P �0.616 for M694V, V726A, and M680I, respectively, inArmenian patients).

To determine whether this departure in the dis-tribution of I/I, I/NI, and NI/NI genotypes from Hardy-Weinberg equilibrium was specific to the Karabakhianand Armenian populations of FMF patients or was alsopresent among the other affected populations, we used asimilar approach to test the distribution of MEFV geno-types in published series of relatively homogeneouspopulations of patients (11–13). Strikingly, in all testedseries, the observed distribution of genotypes differeddramatically from Hardy-Weinberg expectations (P �0.00001) (Figure 2). Importantly, in each of these pop-ulations of different ethnic origins, the disequilibriumwas due to an important paucity of patients with the I/NIgenotype and, conversely, a relative excess of patientswith no identified mutation and with 2 characterizedMEFV mutations, although to a lesser extent.

This finding reveals the existence of an FMF-likecondition that is unrelated to the MEFV gene (seeDiscussion). Thus, we were prompted to estimate thenumber of patients whose phenotype did not result frommutations in the MEFV gene (see Patients and Meth-ods). Our calculations showed that, depending on the

Figure 2. Comparison of the observed distribution of the I/I, I/NI, andNI/NI MEFV genotypes with the theoretical proportion expected fromHardy-Weinberg equilibrium in the Karabakhian patients and inclassically affected populations (Armenians [ref. 9 and the presentstudy], Sephardic Jews [11], Turks [12], and Arabs [13]. Numbers at thetops of the bars are the number of patients in each group. I �identified mutation; NI � nonidentified mutation; Ob � observeddistribution; HW � Hardy-Weinberg equilibrium (binomial distribu-tion), as deduced from the observed frequency of identified andnonidentified MEFV gene mutations. � � P � 0.0016; �� � P �0.00001.

Figure 1. Comparison of the spectrum of MEFV mutations amongArmenian and Karabakhian patients with familial Mediterraneanfever. Frequencies of A, nonidentified (NI) alleles and B, the mostcommon mutations were compared by chi-square test (using Yates’adjustment, when appropriate). Numbers within the bars are thenumber of patients in each group. NS � not significant. � � P � 0.04;�� � P � 0.01; ��� � P � 0.0003; ���� � P � 0.00005.

FMF IN PATIENTS OF KARABAKH AND MEFV GENE ANALYSIS 2327

Page 5: Familial Mediterranean fever among patients from Karabakh and the diagnostic value of MEFV gene analysis in all classically affected populations

series being evaluated, this proportion varied from 7%to 21%, reaching 14% among Karabakhian patients(Table 3). Moreover, the proportion of patients with anNI/NI MEFV genotype whose phenotype could not beexplained by mutations in the MEFV gene ranged from85% among Turkish patients to 99% among Arabpatients (Table 3).

DISCUSSION

In this study, which was designed to further testthe diagnostic value of MEFV gene analysis in patientswith FMF, we evaluated a cohort of patients fromKarabakh and analyzed the genetic data obtained inpopulations classically affected by FMF. In all cases, weobserved a departure from Hardy-Weinberg equilibriumin the MEFV genotype distribution, which resultedessentially from a paucity of patients with only 1 char-acterized mutation and an excess of patients with noidentified mutation. This finding, in combination withother population genetics–based data, reveals the exis-tence of an FMF-like condition that is unrelated to theMEFV gene, and sheds new light on the diagnostic valueof MEFV analysis.

Delineation of the MEFV mutation spectrum inKarabakhian patients revealed a high percentage of NIalleles (26%), as compared with that reported amongArmenian patients (7%) (9). Since no particular MEFVhaplotype was overrepresented among the NI alleles(data not shown), we excluded the existence of a uniqueand rare MEFV mutation resulting from a founder

effect, which would have escaped our mutation screen-ing procedure. When considering the characterized mu-tated alleles, 3 of the 6 MEFV mutations identified inKarabakhian patients (i.e., M694V, V726A, and R761H)were found at different frequencies in the Armenianpatients. Together, these results show that the MEFVmutation spectrum differs significantly between these 2populations living in neighboring areas, an observationthat may result from a founder effect and the relativelyindependent evolution of each population since thebeginning of the twentieth century.

Most important, the distribution of I/I, I/NI, andNI/NI genotypes at the MEFV locus differed dramati-cally from a binomial distribution among Karabakhianpatients, as well as among patients of Armenian, Turk-ish, Arab, or Sephardic Jewish ancestry (Figure 2). Atleast 3 hypotheses may account for this observation:frequent consanguineous unions in these populations, apossible recruitment bias, and the existence of an FMFphenotype unrelated to MEFV in several patients withno identified mutation.

Many lines of evidence strongly oppose the first 2hypotheses. There was no known consanguinity in theArmenian and Karabakhian patient populations. Mostimportant, the distribution of the most common MEFVmutations in these 2 populations did not differ signifi-cantly from Hardy-Weinberg expectations. This obser-vation is evidence against a selection bias that wouldhave favored the recruitment of severely affected pa-tients into the study. It also argues against the hypothesis

Table 3. Estimate of the number of patients with FMF from Karabakh and from classically affected populations whose phenotype results from ordoes not result from mutations in the MEFV gene*

Origin of thepatients Author, year (ref.) nI/I nI/NI nNI/NI NTotal NMEFV NOther

NOther NOther

NTotal nNI/NI

Karabakhian Present study 30 8 7 45 38.5 6.5 0.14 0.93Armenian Cazeneuve et al, 1999 (9)

and present study†126 11 10 147 137.2 9.8 0.07 0.98

Sephardic Jewish Livneh et al, 1999 (11) 133 31 14 178 165.8 12.2 0.07 0.87Turkish Akar et al, 1999 (12) 139 55 36 230 199.4 30.6 0.13 0.85Arab Shinawi et al, 2000 (13) 45 6 14 65 51.2 13.8 0.21 0.99

* Calculations were performed under the assumption that the Hardy-Weinberg requirements were fulfilled. With regard to the absence ofconsanguinity, this information is available only for the Karabakhian and Armenian populations. nI/I � number of patients carrying an I/I genotype;nI/NI � number of patients carrying an I/NI genotype; nNI/NI � number of patients carrying an NI/NI genotype; NTotal � total number of patientsin the series; NMEFV � number of patients whose disease phenotype is due to mutations in the MEFV gene, i.e., under our hypothesis,

NMEFV ��nI/I�

nI/NI

2 �2

nI/I

(see Patients and Methods for details); NOther � number of patients whose phenotype does not result from mutations in MEFV (NOther � NTotal� NMEFV).† These data include a series of 85 patients described previously (9) and 41 Armenian patients with FMF evaluated in the present study.

2328 CAZENEUVE ET AL

Page 6: Familial Mediterranean fever among patients from Karabakh and the diagnostic value of MEFV gene analysis in all classically affected populations

of an increased embryonic death of zygotes with 2 severeMEFV mutations; indeed, under such a hypothesis, thenumber of patients carrying these mutations would bemuch lower than would be expected from Hardy-Weinberg equilibrium, a situation reported in otherconditions, such as congenital disorders of glycosylationtype Ia (14). The binomial distribution of the mostcommon MEFV mutations, therefore, reflects randominclusion of FMF patients in the above-mentioned stud-ies and further supports the fact that in these populationsamples, the requirements for Hardy-Weinberg equili-brium were indeed fulfilled.

It is important to emphasize that the deviationsfrom Hardy-Weinberg equilibrium documented in allthe population samples evaluated herein cannot beexplained by the existence of unidentified MEFV muta-tions that would have escaped the mutation search.Indeed, in this latter situation, whatever the moleculardiagnostic procedure (i.e., complete screening for allMEFV exons [ref. 9 and the present study] or search fora limited number of known mutations [11–13]), the NIalleles should be normally distributed. This was not thecase (Figure 2). Taken together, these data allow us toreject the first two hypotheses. We therefore concludethat, according to the third hypothesis, the departurefrom Hardy-Weinberg equilibrium documented in thesepopulations of FMF patients reveals the existence of anFMF-like phenotype that is unrelated to MEFV inseveral patients with no identified MEFV mutation.

Although the patients evaluated in each seriesfulfilled the established clinical criteria for a diagnosis ofFMF, the implication of MEFV in their disease pheno-type can be ascertained only in those who had 2 identi-fied MEFV mutations. Most of the remaining patientsmay therefore have other conditions, including condi-tions that possibly result from mutations at other loci.The family history of FMF among the relatives of several

Karabakhian patients who had no identified MEFVmutation (data not shown) further supports the latterhypothesis. The existence of a locus heterogeneity inFMF has already been suggested from the findings ofevaluations of 2 Turkish families (15). However, thathypothesis was based on a segregation analysis of poly-morphic markers of the 16p13 region, and not on theanalysis of the MEFV gene, which was unknown at thattime. Under these experimental conditions, an incom-plete penetrance of the disease phenotype as well asintrafamilial MEFV allelic heterogeneity—two well-known features of FMF that have been recently docu-mented in all classically affected populations (9,16–19)—could easily account for the absence of linkagebetween those markers and the disease.

Having shown that the Hardy-Weinberg disequili-brium documented in the present study indeed reflectsthe existence of a group of patients whose phenotypedoes not result from mutations in the MEFV gene, weestimated the number of these patients. Depending onthe patients’ ancestry, the estimated proportion of NI/NIpatients whose phenotype did not result from MEFVmutations was found to vary from 85% to 99% (i.e.,7–21% of the patients with a clinical diagnosis of FMF).It must be kept in mind that in each series, this propor-tion is underestimated. Indeed, given the high frequencyof healthy individuals with an MEFV mutation in theheterozygous state in the 4 at-risk populations(13,16,20,21), the phenotype of several patients with anI/NI genotype at the MEFV locus probably does notresult from mutations in the MEFV gene.

What are the consequences of these findings forthe diagnosis of FMF? First, our results should beconsidered in light of the efficiency of the molecularstrategies for detecting 2 mutated MEFV alleles thatwere used by the different groups of investigators.Although the MEFV mutation spectrum differs among

Table 4. Diagnostic value of MEFV gene analysis in the Karabakhian population of patients and in classically affected populations

Origin of thepatients Author, year (ref.) Diagnostic strategy

% of patients

Two identifiedMEFV mutations (I/I)

No identifiedMEFV mutations (NI/NI)

Karabakhian Present study DGGE* 66.7 15.6Armenian Cazeneuve et al, 1999 (9) DGGE* 89.4 3.5Sephardic Jewish Livneh et al, 1999 (11) Search for 4 mutations† 74.7 7.9Turkish Akar et al, 1999 (12) Search for 7 mutations‡ 60.4 15.7Arab Shinawi et al, 2000 (13) Search for 4 mutations§ 69.2 21.5

* Denaturing gradient gel electrophoresis (DGGE) was designed to search for MEFV mutations in all coding sequences and intron boundaries.† Mutations M694V, V726A, M680I, and E148Q were sought.‡ Mutations M694V, V726A, M680I, M694I, R761H, K695R, and E148Q were sought.§ Mutations M694V, V726A, M680I, and M694I were sought.

FMF IN PATIENTS OF KARABAKH AND MEFV GENE ANALYSIS 2329

Page 7: Familial Mediterranean fever among patients from Karabakh and the diagnostic value of MEFV gene analysis in all classically affected populations

classically affected populations, the routine diagnosticprocedure is based on a search for a limited number ofknown mutations. As shown in Table 4, this strategyleaves 7.9–21.5% of the patients with 2 uncharacterizedFMF alleles, depending on the ancestry of the patients.In this regard, the complete screening of all codingsequences and intron boundaries performed in the Ar-menian patients appears to be particularly efficient,since the 2 FMF alleles remained uncharacterized inonly 3.5% of those patients (9).

At first glance, this observation may encouragethe use of diagnostic procedures based on the analysis ofall 10 MEFV exons. However, using this strategy, weshowed that up to 15.6% of the Karabakhian patientshad no identified MEFV mutation. Furthermore, the aposteriori analysis of the MEFV mutation spectrum inthe Armenian (9) and Karabakhian populations indi-cates that the search for 6 mutations (i.e., M694V,V726A, M680I, F479L, R761H, and E148Q) would haveprovided a similar rate of identified mutations. Takentogether, these data do not support the time-consumingand costly diagnostic strategy based on a completescreening of all coding sequences and intron boundaries.Second, and most important, in each affected popula-tion, the estimation of the number of patients whosephenotype did not result from mutations in MEFV(NOther) allowed us to calculate the proportion of suchpatients among individuals with an NI/NI genotype(Table 3). Strikingly, whatever the ancestry of the pop-ulations and whatever the diagnostic procedure, thisratio (NOther/nNI/NI) is particularly high, ranging from0.85 to 0.99. The proportion reaches 0.99 in the Arabpopulation; such a high proportion in this population isall the more since striking, first, only 69.2% of thepatients carried mutated MEFV alleles—as identified bythe search for only 4 mutations (Table 4)—and second,according to our calculations, at least 21% of thesepatients presented with an FMF-like phenotype that didnot result from mutations in MEFV (Table 3).

From a more general viewpoint, the conse-quences of this study are 2-fold. The complete screeningof the MEFV gene in patients presenting with a clinicaldiagnosis of FMF (according to the established clinicalcriteria [4]) is expected to identify only a few patientswith 2 mutated alleles that would not have been identi-fied by means of a diagnostic procedure designed tosearch for a limited number of known mutations. Inother words, the probability that a patient indeed carries2 MEFV mutations despite the negative results of alimited search for mutations is particularly low, rangingfrom 1% to 15%, depending on the patient’s ethnic

background (Table 3). Thus, the present study disclosesthe existence of an FMF-like disorder in all classicallyaffected populations that is not due to mutations in theMEFV gene. The identification of the pathophysiologicmechanisms underlying this newly recognized disorderwill be essential in clarifying the nosology of a hetero-geneous group of diseases that, to date, are phenotypi-cally indiscernible.

ACKNOWLEDGMENTS

We are indebted to the patients and their families fortheir willingness to participate in this study.

REFERENCES

1. The French FMF Consortium. A candidate gene for familialMediterranean fever. Nat Genet 1997;17:25–31.

2. The International FMF Consortium. Ancient missense mutationsin a new member of the RoRet gene family are likely to causefamilial Mediterranean fever. Cell 1997;90:797–807.

3. Sohar E, Gafni J, Pras M, Heller H. Familial Mediterranean fever:a survey of 470 cases and review of the literature. Am J Med1967;43:227–53.

4. Livneh A, Langevitz P, Zemer D, Zaks N, Kees S, Lidar T, et al.Criteria for the diagnosis of familial Mediterranean fever. ArthritisRheum 1997;40:1879–85.

5. Dinarello CA, Wolff SM, Goldfinger SE, Dale DC, Alling DW.Colchicine therapy for familial Mediterranean fever: a double-blind trial. N Engl J Med 1974;291:934–7.

6. Zemer D, Revach M, Pras M, Modan B, Schor S, Sohar E, et al.A controlled trial of colchicine in preventing attacks of familialMediterranean fever. N Engl J Med 1974;291:932–4.

7. Zemer D, Pras M, Sohar E, Modan M, Cabili S, Gafni J.Colchicine in the prevention and treatment of the amyloidosis offamilial Mediterranean fever. N Engl J Med 1986;314:1001–5.

8. Cazeneuve C, Ajrapetyan H, Papin S, Roudot-Thoraval F, Genev-ieve D, Mndjoyan E, et al. Identification of MEFV-independentmodifying genetic factors for familial Mediterranean fever. Am JHum Genet 2000;67:1136–43.

9. Cazeneuve C, Sarkisian T, Pecheux C, Dervichian M, Nedelec B,Reinert P, et al. MEFV-gene analysis in Armenian patients withfamilial Mediterranean fever: diagnostic value and unfavorablerenal prognosis of the M694V homozygous genotype—genetic andtherapeutic implications. Am J Hum Genet 1999;65:88–97.

10. Cotton RG. Slowly but surely toward better scanning for muta-tions. Trends Genet 1997;13:43–6.

11. Livneh A, Langevitz P, Shinar Y, Zaks N, Kastner DL, Pras M, etal. MEFV mutation analysis in patients suffering from amyloidosisof familial Mediterranean fever. Amyloid 1999;6:1–6.

12. Akar N, Misiroglu M, Yalcinkaya F, Akar E, Cakar N, Tumer N,et al. MEFV mutations in Turkish patients suffering from familialMediterranean fever. Hum Mutat 2000;15:118–99.

13. Shinawi M, Brik R, Berant M, Kasinetz L, Gershoni-Baruch R.Familial Mediterranean fever: high gene frequency and hetero-geneous disease among an Israeli-Arab population. J Rheumatol2000;27:1492–5.

14. Schollen E, Kjaergaard S, Legius E, Schwartz M, Matthijs G. Lackof Hardy-Weinberg equilibrium for the most prevalent PMM2mutation in CDG-Ia (congenital disorders of glycosylation typeIa). Eur J Hum Genet 2000;8:367–71.

15. Akarsu A, Saatci U, Ozen S, Bakkaloglu A, Besbas N, Sarfarazi M.Genetic linkage study of familial Mediterranean fever (FMF) to

2330 CAZENEUVE ET AL

Page 8: Familial Mediterranean fever among patients from Karabakh and the diagnostic value of MEFV gene analysis in all classically affected populations

16p13.3 and evidence for genetic heterogeneity in the Turkishpopulation. J Med Genet 1997;34:573–8.

16. Yilmaz E, Ozen S, Balci B, Duzova A, Topaloglu R, Besbas N, etal. Mutation frequency of familial Mediterranean fever and evi-dence for a high carrier rate in the Turkish population. Eur J HumGenet 2001;9:553–5.

17. Aksentijevich I, Torosyan Y, Samuels J, Centola M, Pras E, ChaeJJ, et al. Mutations and haplotype studies of familial Mediterra-nean fever reveal new ancestral relationships and evidence for ahigh carrier frequency with reduced penetrance in the AshkenaziJewish population. Am J Hum Genet 1999;64:949–62.

18. Gershoni-Baruch R, Shinawi M, Leah K, Badarnah K, Brik R.Familial Mediterranean fever: prevalence, penetrance and geneticdrift. Eur J Hum Genet 2001;9:634–7.

19. Gershoni-Baruch R, Shinawi M, Shamaly H, Katsinetz L, Brik R.Familial Mediterranean fever: the segregation of 4 differentmutations in 13 individuals from one inbred family: genotype-phenotype correlation and intrafamilial variability. Am J MedGenet 2002;109:198–201.

20. Torosyan Y, Aksentijevich I, Sarkisian T, Astvatsatryan V, Ay-vazyan A, Centola M, et al. A population-based survey reveals anextremely high FMF carrier frequency in Armenia, suggestingheterozygote advantage [abstract]. Am J Hum Genet 1999;65Suppl:A400.

21. Stoffman N, Magal N, Shohat T, Lotan R, Koman S, Oron A, et al.Higher than expected carrier rates for familial Mediterraneanfever in various Jewish ethnic groups. Eur J Hum Genet 2000;8:307–10.

FMF IN PATIENTS OF KARABAKH AND MEFV GENE ANALYSIS 2331