Bobadilla 2002

HUMAN MUTATION 19:575606 (2002)
REVIEW ARTICLE
Cystic Fibrosis: A Worldwide Analysis of CFTR

MutationsCorrelation With Incidence Data and
Application to Screening
Joseph L. Bobadilla,1 Milan Macek Jr.,2 Jason P. Fine,3 and Philip M. Farrell1*
1
Department of Pediatrics, University of Wisconsin, Madison, Wisconsin
2
Institute of Biology and Medical GeneticsCystic Fibrosis Center, Charles University, Prague, Czech Republic
3
Department of Biostatistics and Medical Informatics, University of Wisconsin, Madison, Wisconsin
Communicated by Lap-Chee Tsui
Although there have been numerous reports from around the world of mutations in the gene of
chromosome 7 known as CFTR (cystic fibrosis transmembrane conductance regulator), little at-
tention has been given to integrating these mutant alleles into a global understanding of the popu-
lation molecular genetics associated with cystic fibrosis (CF). We determined the distribution of
CFTR mutations in as many regions throughout the world as possible in an effort designed to: 1)
increase our understanding of ancestrygenotype relationships, 2) compare mutational arrays with
disease incidence, and 3) gain insight for decisions regarding screening program enhancement
through CFTR multi-mutational analyses. Information on all mutations that have been published
since the identification and cloning of the CFTR genes most common allele, DF508 (or F508del),
was reviewed and integrated into a centralized database. The data were then sorted and regional
CFTR arrays were determined using mutations that appeared in a given region with a frequency of
0.5% or greater. Final analyses were based on 72,431 CF chromosomes, using data compiled from
over 100 original papers, and over 80 regions from around the world, including all nations where
CF has been studied using analytical molecular genetics. Initial results confirmed wide mutational
heterogeneity throughout the world; however, characterization of the most common mutations
across most populations was possible. We also examined CF incidence, DF508 frequency, and
regional mutational heterogeneity in a subset of populations. Data for these analyses were filtered
for reliability and methodological strength before being incorporated into the final analysis. Statis-
tical assessment of these variables revealed that there is a significant positive correlation between
DF508 frequency and the CF incidence levels of regional populations. Regional analyses were also
performed to search for trends in the distribution of CFTR mutations across migrant and related
populations; this led to clarification of ancestrygenotype patterns that can be used to design
CFTR multi-mutation panels for CF screening programs. From comprehensive assessment of these
data, we offer recommendations that multiple CFTR alleles should eventually be included to in-
crease the sensitivity of newborn screening programs employing two-tier testing with trypsinogen
and DNA analysis. Hum Mutat 19:575606, 2002. © 2002 Wiley-Liss, Inc.
KEY WORDS: cystic fibrosis transmembrane conductance regulator; CFTR; cystic fibrosis; CF; newborn;
mutation screening; incidence; ABCC7
Received 19 June 2001; accepted revised manuscript 1 March Chapter; Contract grant sponsor: National Institutes of Health;
2001. Contract grant numbers: DK 34108; M01 RR03186; Contract
*Correspondence to: Philip M. Farrell, M.D., Ph.D., Professor grant sponsor: Internal Grant Agency of the Czech Ministry of
Health (IGA MZ CR); Contract grant number: 6250-3; 6464-3;
of Pediatrics and Dean, University of Wisconsin Medical School,
00000064203; Contract grant sponsor: Ministry of Education of
Room 1217 MSC, 1300 University Avenue, Madison, WI 53706-
1532. E-mail: [email protected] the Czech Republic; Contract grant numbers: 111300003; LN
00A079; ME451.
Contract grant sponsors: National Center for Research Re-
sources to the University of Wisconsin Medical School; Academic DOI: 10.1002/humu.10041
Generalist Scholarship Program from the American College of Published online in Wiley InterScience (www.interscience.wiley.
Physicians-American Society of Internal Medicine, Wisconsin com).
© 2002 Wiley-Liss, Inc.

576 BOBADILLA ET AL.
DATABASES:
CFTR OMIM: 602421; 219700 (CF); GDB: 120584; XM_004980;
http://www.genet.sickkids.on.ca/cftr/ (CFTR Database; HGMD: CFTR)
INTRODUCTION 1994]. The large number of more common mu-

Cystic fibrosis (CF; MIM# 219700) is one of tations and the association of certain alleles with
the most common life-threatening autosomal re- ethnic populations, such as African Americans
cessive disorders in the Western World [Welsh et [Macek et al., 1997], has presented challenges
al., 1995; Farrell, 2000]. Previous investigators for screening programs attempting to optimize
have generated evidence that CF-causing muta- sensitivity and encompass minority populations.
tions have existed since the Paleolithic period In a recent statement publication, the American
(50,000 years) [Morral et al., 1994], and many College of Medical Genetics (ACGM) has iden-
are strongly associated with European-derived tified several issues including the differing preva-
populations. CF is caused by mutations in the lence of individual CFTR mutations among
cystic fibrosis transmembrane conductance regu- various ethnic, demographic, and racial groups;
lator gene (CFTR; MIM# 602421), discovered the extreme ethnic heterogeneity of the U.S.
in 1989 [Riordan et al., 1989; Rommens et al., population and its increasing admixture; and the
1989]. The most common mutant allele is the absence of guidelines for developing an appropri-
F508del (∆F508) mutation [Kerem et al., 1989], ate mutation test panel [Grody et al., 2001].
which is a three-nucleotide deletion at the 508th According to the literature on European-de-
codon causing the deletion of a phenylalanine rived populations, CF is usually estimated to af-
residue and subsequent defective intracellular fect as many as 1 in 2,0003,000 Caucasian
processing of the CFTR protein that is an impor- newborns [Farrell, 2000], with a carrier fre-
tant chloride channel [Welsh et al., 1995; Mo- quency that averages out to a striking 1 in 26
rales et al., 1999]. Worldwide, the ∆F508 individuals in such populations [Romeo et al.,
mutation is responsible for approximately two- 1989; Welsh et al., 1995]. Data on derivative
thirds (66%) of all CF chromosomes; however, populations in North America indicate a lower
there is great mutational heterogeneity in the re- incidence, namely about 1:3,500 [Kosorok et al.,
maining one-third of all alleles [Cystic Fibrosis 1996]. However, CF is still the most prevalent
Genetic Analysis Consortium (CFGAC), 1994; hereditary metabolic disorder detected through
Estivill et al., 1997; Welsh et al., 1995]. Nearly newborn screening [Allen and Farrell, 1996]. CF
1,000 CFTR alleles have been identified to date neonatal screening, as a method of early diag-
[Cystic Fibrosis Genetic Analysis Consortium nosis, has been implemented in many regions
(CFGAC), 2000] (www.genet.sickkids.on.ca/ using a two-tier method of immunoreactive
cftr). These mutations have been classified into trypsinogen (IRT) analysis first and then, if the
five categories depending on their functional ef- level is high, DNA analysis either for the ∆ F508
fects vis-à-vis the CFTR protein [Zielenski and mutation or multiple CFTR alleles [Farrell,
Tsui, 1995]. Although these mutations vary 2000]. Despite the attractiveness of this diag-
greatly in their frequency and distribution, the nostic strategy, neonatal screening using the two-
vast majority are either private or limited to a tier, IRT/DNA procedure [Gregg et al., 1997]
small number of individuals. A group of CFTR cannot achieve 100% sensitivity in most regions
alleles also exists in non-Caucasian populations when the DNA interrogation is limited to the
[Macek et al., 1997]. There are about 10 to 20 ∆F508 allele.
less common mutations that exist at over 0.1% The purpose of this investigation was two-
worldwide. Some mutations can reach higher fre- fold. First, a detailed map of the CFTR muta-
quencies in particular populations due to founder tional spectrum across the world was created by
effect (e.g., M1101K in Hutterites) [Zielenski et systematic literature survey. Once these muta-
al., 1993] or genetic drift (e.g., G551D in the his- tional arrays were synthesized, trends were
toric Celts), and other mutations are recurrent searched for and analyzed on an ethnic, regional,
(e.g., R1162X in Zuni and Northwest Italian and global scale. Secondly, we addressed the
populations) which explains their presence in his- impact of ∆F508 on mutational heterogeneity
torically unrelated populations [Morral et al., and the incidence of classical CF. This was done
WORLDWIDE ANALYSIS OF CFTR MUTATIONS 577
in an attempt to quantify a pattern between the percentages of the total number of chromosomes
number of mutations present in a population and in the associated studies. These results appear
the CF incidence and/or the ∆F508 frequency in parentheses next to the mutations for each
of that population. Finding a trend and devel- region. For those studies that had fewer than 50
oping an algorithm with these variables could chromosomes (nc<50), but more than 20
allow for the estimation of CF incidence and (nc>20), mutations were ranked numerically in
mutational array size in regions where current the order of most common to least common and
data sets are incomplete. Together, the results rough estimate percentages appear in parenthe-
of these two investigations should provide help- ses. For those studies that had fewer than 20
ful information for those regions that know the chromosomes (nc<20), mutations were only
ethnic origin of their population and wish to ranked in the order of most common to least
implement CF screening programs. Indeed, the common; this was done to prevent percentage
exhaustive survey and analyses we performed can skewing that lower n studies are prone to.
serve the need for data allowing determination When data from multiple studies were appli-
of an appropriate mutation panel [Grody et al., cable to the same population or geographic re-
2001] that could be used in regional CF neona- gion, several steps were taken to prevent biasing
tal screening programs. of the true mutation frequencies. First, attempts
were made to determine if there was overlap in
METHODS AND CRITERIA
the populations studied. If there was identifiable
This study consisted of two components. overlap, the study that probed for the greatest
Phase I consisted of identification and geo- number of mutations was used and the other
graphic mapping of the mutational arrays for discarded, unless significant discrepancies or
various countries and regions across the world. novel mutations were noted in the reported re-
Phase II consisted of searching for global trends sults. If there was no overtly detectable overlap,
in mutation distribution as well as comparison the methods used for mutational analysis were
of published incidence data with the CFTR reviewed. If a specific technique used was not
mutational distribution patterns. In order to pre- able to detect a certain mutation found in a sec-
serve and enhance data integrity during phase ond study, the missing mutation was simply
II, a two-tier filtering protocol was developed. added to the final list with the percent found in
the second study unmodified. If a mutation was
Mutational Array Determination
found at different frequencies in different stud-
Phase I consisted of a comprehensive litera- ies, the frequencies were averaged to compute a
ture search, dating back to the identification and final mutational prevalence. Finally, if a muta-
cloning of the CFTR gene in 1989 [Riordan et tion was found in one study, but not another,
at., 1989; Rommens et al., 1989; Kerem et al., even though the methods used allowed for its
1989] and extending until December 31, 2000. detection in both, a 0% was added into the av-
The OVID MedLine database was used to gather eraging equation. By using these techniques in
initial sources for review. The search protocol calculating the final mutational frequencies,
used an initial query of Cystic Fibrosis with an potential errors were minimized. In preparing the
AND Boolean string to include the following final mutational arrays for each country or re-
subheadings: Ethnology OR Epidemiology gion, a cutoff value of 0.5% was used as a lower
OR Etiology OR Genetics OR Diagnosis. limit. This threshold was utilized to minimize
All articles identified by these search parameters the contribution of private mutations and se-
underwent an abstract review, and those perti- lect for the actual analysis of those mutations
nent to this study were selected for further criti- that are common throughout distinct popula-
cal review. All papers that were identified by tions and regions. We focused our study on clas-
these combined search methods were critically sical CF to avoid other conditions associated
reviewed, and relevant data were pooled into a with CFTR mutations.
centralized database for further analysis. With After completing the analyses for 19892000
regard to the final data presented in this paper, in the initial manuscript, we performed a simi-
studies that included greater than 50 CF chro- lar literature survey for calendar year 2001 and
mosomes (nc>50) were represented as mutation identified 17 more articles. These were exam-
ined thoroughly and found to be confirmatory used for analysis, and also to eliminate any in-
of our conclusions, although many of them were ternal methodological biases. Filter 1 was de-
for isolated regions or included only a few pa- signed to remove weak data sets. In order to pass
tients, while others reported population screen- through this filter, each country study was re-
ing information (rather than CFTR mutational quired to meet the following three criteria: 1) a
assessment of aggregated CF patients). Most sig- reliable ∆F508 percentage (nc>50), 2) a pub-
nificantly, the quantitative CFTR data con- lished CF incidence value, and 3) a CF muta-
firmed our findings and recommendations for tion array large enough to generate percentages
Europe and North America. The implications of total CF alleles (determined in phase I,
of the relevant articles are described in either nc>50). The goal of this filter was to remove
Table 1 or the text. In addition, the Central and regions with incomplete data sets and lower n
Eastern European data presented in Table 1 were studies. This filter was used when linear regres-
available either from abstracts or the collabora- sion analysis was done with regard to ∆F508 per-
tive research of one of the authors (M.M.). centage versus incidence. Filter 2 was designed
to remove those studies that were using allele
Statistical Methodology
specific mutation detection methods that could
Phase II of the study involved an examina- bias the results. During this filter implementa-
tion of trends in mutation distribution as a function, the original methods of mutation detec-
tion of ethnic and geographic factors. We also tion were reviewed. Only those studies that used
assessed potential correlations of the mutation a gene scanning technique (e.g., single strand
arrays determined in phase I with the CF inci- conformation polymorphism (SSCP), denatur-
dence levels of all countries that had such data ing gradient gel electrophoresis (DGGE), etc.),
reliably determined. Comparisons of CFTR al- or those that probed for greater than 200 muta-
lele percentages and CF diagnoses in different tions by comprehensive allele specific methods,
populations were based on standard two sample were selected to pass through filter 2. The goal
tests for equal proportions. This includes the of this filter was to prevent the downward skew-
analyses of Lac St. Jean-Quebec-Toronto, ∆F508 ing of mutational arrays size by selecting only
in South America, and G542X in Brittany and those studies that could, in theory, detect 100%
Southern France. Comparisons of average num- of the present CFTR mutations, or a large por-
ber of mutations in Bulgaria/Spain/Turkey/ tion thereof. When doing linear regression analy-
Greece versus other European countries used the sis of the total number of mutations present at
Wilcoxon rank sum test. Comparison of percent greater than 0.5% versus ∆F508 percentage or
diagnoses in these regions used an exact rank CF incidence, both filters 1 and 2 were employed
sum test. These tests are well suited to the small and both ∆F508 percentage and CF incidence
number of countries investigated in these mini- were log transformed (Fig. 1).
analyses. The potential relationships between
incidence, percentage ∆F508, and the number HISTORICAL PERSPECTIVES AND
of mutations were evaluated via linear regres- STATISTICAL IMPLICATIONS
sion. The analyses involving number of muta- Worldwide, the ∆F508 allele, which presum-
tions were restricted to those countries whose ably arose from a single origin, ranks as the most
data satisfied the inclusion criterion listed for common CFTR mutation [Cystic Fibrosis Ge-
the two filters as described below. Furthermore, netic Analysis Consortium (CFGAC), 1994].
when comparing number of mutations to ∆F508 Although ∆F508 is the most prevalent mutation
percentage and CF incidence, these latter two for most populations, there are some, such as
variables were log transformed to help normal- certain Jewish groups and many Middle Eastern
ize the data. Comparison of ∆F508 percentage populations, in which other mutations show a
to CF incidence was done using only filter 1, as higher frequency (Table 1). Examining the
there was no need to control for mutational ar- ∆F508 allele across European countries illus-
ray size with this analysis. trates how variable the CFTR mutation frequen-
Two filters were employed during the linear cies are as ancestry varies. Indeed, ∆F508
regression analysis of phase II. These filters were frequencies vary from a maximum of 100% in
designed to ensure the integrity of the data set the isolated Faroe Islands of Denmark [Schwartz
FIGURE 1. Flow chart of study methods. This figure graphically illustrates the methods used to sort and analyze the
individual studies incorporated into this work. Filter 1 is marked by a country fulfilling all of the following three criteria:
1) having and available ∆F508 percentage; 2) having a published CF incidence value; and 3) having an nc>50. Filter 2
is marked by fulfilling the following criteria: 1) passing filter 1; and 2) mutation identification via a gene scanning methods
(DGGE, SSCP, etc.), or probing greater than 200 allele specific mutation loci.
et al., 1995a] to a minimum of 24.5% in Turkey 1997]. Mutations that fall into this category in-
(Table 1). There is a clear northwest to south- clude: 1) G542X, of single origin, associated with
east gradient in ∆F508 frequency across Europe the ancient Phoenicians [Loirat et al., 1997], 2)
[European Working Group on CF Genetics, N1303K, also of single origin, and linked to an-
1990; Lucotte et al., 1995]. cient Mediterranean populations, and 3) G551D,
There are several other CFTR mutations that also of single origin [Cashman et al., 1995], hav-
are common throughout the entire world. His- ing been associated with the ancient Celtic tribes
torically, these alleles appear to be the oldest, [Macek et al., 1991].
other than ∆F508, and therefore have had the There are also a multitude of other interest-
longest time to spread throughout different ing regions or populations such as Reunion Is-
populations [European Working Group on CF land, which shows its distinct ancestral founder
Genetics, 1990; Morral et al., 1994; Loirat et al., effect by having both high rates of G551D and
580
TABLE 1. Mutational Arrays, Detection Rates and Methods by Region*
Estimated
Projected detection of Number of Number of
Country/ allele two CFTR mutations chromosomes
detectiona mutationsb includedc (max/min)d
BOBADILLA ET AL.
Region Mutation array Reference
Europe
Albania ∆F508 (72.4%) C276X (0.7%) 74.5 55.5 4 270/146 CFGAC [1994]; Macek et al.
G85E (0.7%) R1070Q (0.7%) [2002]
Austria ∆F508 (62.9%) 457TAT→G (1.2%) 76.6 58.7 11 1516/580 Estiville et al. [1997]; Dörk et al.
(total) G542X (3.3%) 2183AA→G (0.7%) [2000]; Macek et al. [2002]
CFTRdele2,3 (2.1%) N1303K (0.6%)
R1162X (1.9%) I148T (0.5%)
R553X (1.7%) R117H (0.5%)
G551D (1.2%)
Austria ∆F508 (74.6%) 2183AA→G (2.4%) 95.3 90.8 8 126 Stuhrmann et al. [1997]
(tyrol) R1162X (8.7%) G551D (1.6%)
G542X (2.4%) R347P (1.6%)
2789+5G→A (2.4%) Q39X (1.6%)
Belarus ∆F508 (61.2%) R553X (0.5%) 75.2 56.6 9 278/188 Dörk et al. [2000]; Macek et al.
G542X (4.5%) R334W (0.5%) [2002]
CFTRdele2,3 (3.3%) R347P (0.5%)
N1303K (3.2%) S549N (0.5%)
W1282X (1.0%)
Belgium ∆F508 (75.1%) 622-1A→C (0.5%) 100.0 100.0 27 1504/522 Cuppens et al. [1993]; Mercier et
G542X (3.5%) G458V (0.5%) al. [1993]; CFGAC [1994];
N1303K (2.7%) 1898+G→C (0.5%) Estivill et al.[1997]
R553X (1.7%) G970R (0.5%)
1717-1G→A (1.6%) 4218insT (0.5%)
E60X (1.6%) 394delTT (0.5%)
W1282X (1.4%) K830X (0.5%)
2183A→G+2184delA (1.2%) E822K (0.5%)
W401X (1.0%) 3272-1G→A (0.5%)
A455E (1.0%) S1161R (0.5%)
3272-26A→G (1.0%) R1162X (0.5%)
S1251N (1.0%) 3750delAG (0.5%)
S1235R (0.8%) S1255P (0.5%)
∆I507 (0.6%)
Bulgaria ∆F508 (63.6%) R75Q (1.0%) 93.0 86.5 21 948/432 Angelicheva et al. [1997];
(total) N1303K (5.6%) 2183AA→G (0.9%) Estivill et al. [1997]; Macek
G542X (3.9%) G1244V+S912L (0.9%) et al. [2002]
R347P (2.2%) G85E (0.9%)
1677delTA (2.1%) 2184insA (0.9%)
R1070Q (1.8%) L88X+G1069R (0.8%)
Q220X (1.2%) 2789+5G→A (0.8%)
3849+10KbC→T (1.1%) G1244E (0.8%)
W1282X (1.0%) 1717-1G→A (0.8%)
2176insC (1.0%) Y919C (0.7%)
G1069R (1.0%)
Bulgaria 1) DF508 4) 1677delTA — — 6 13 Angelicheva et al. [1997]
(ethnic 2) R347P 5) Q493R
Turks) 3) G542X 6) L571S — — 1 30 Angelicheva et al. [1997]
Bulgaria 1) DF508 (100.0%)
(Gypsy)
Croatia ∆F508 (64.5%) G551D (1.1%) 72.5 52.6 5 276 Macek et al. [2002]
G542X (3.3%) 3849+10KbC→T (0.7%)
N1303K (2.9%)
Czech ∆F508 (70.0%) 1898+1G→T (2.0%) 89.6 80.3 10 2196/628 CFGAC [1994]; Estiville et al.
Republic CFTRdele2,3 (5.5%) 2143delT (1.2%) [1997]; Dörk et al. [2000];
G551D (3.8%) R347P (0.8%) Macek et al. [2002]
N1303K (2.9%) 3849+10KbC→T (0.6%)
G542X (2.2%) W1282X (0.6%)
Denmark ∆F508 (87.5%) G542X (0.7%) 92.3 85.2 6 1888/678 CFGAC [1994]; Schwartz et al.
(excluding 394delTT (1.8%) 621+1G→T (0.6%) [1994]; Estiville et al. [1997]
Faroe) N1303K (1.1%) 3659delC (0.6%)
Estonia ∆F508 (51.7%) R117C (1.7%) 80.2 64.3 10 165/80 Estivill et al. [1997]; Klaassen et
394delTT (13.3%) E217G (1.7%) al. [1998]; Macek et al.
S1235R (3.3%) R1066H (1.7%) [2002]
359insT (1.7%) 3659delC (1.7%)
I1005R (1.7%) S1169X (1.7%)
Finland ∆F508 (46.2%) G542X (1.9%) 78.8 62.1 4 132/52 CFGAC [1994]; Kere et al.
394delTT (28.8%) 3372delA (1.9%) [1994]; Estivill et al. [1997]
WORLDWIDE ANALYSIS OF
France ∆F508 (67.7%) 2789+5G→T (0.79%) 79.7 63.6 12 17854/7420 Chevalier-Porst et al. [1994];
(total) G542X (2.94%) 2184delA+2183A→G (0.77%) Estivill et al. [1997]; Claustres
et al. [2000]; Guilloud-Bataille
N1303K (1.83%) G551D (0.74%) et al. [2000]
1717-1G→A (1.35%) 1078delT (0.63%)
W1282X (0.91%) ∆I507 (0.62%)
R553X (0.86%) Y122K (0.59%)
France ∆F508 (75.8%) R297Q (0.8%) 98.7 97.4 18 599/365 Férec et al. [1992]; Scotet et al.
(Brittany) 1078delT (4.0%) R347H (0.8%) [2000]
G551D (3.6%) I1234V (0.8%)
N1303K (3.0%) R553X (0.8%)
R117H (1.7%) 2789+5G→A (0.8%)
3272-26A→G (1.3%) 4005+1G→A (0.7%)
CFTR MUTATIONS
G542X (1.1%) 621+1G→T (0.6%)
1717-1G→A (1.0%) ∆I507 (0.6%)
G1249R (0.8%) W846X (0.5%)
France ∆F508 (70.0%) N1303K (0.8%) 90.4 81.7 16 250 Claustres et al. [1993]
(southern) G542X (6.4%) 3737delA (0.8%)
1717-1G→A (1.6%) R1162X (0.8%)
L206W (1.2%) Y1092X (0.8%)
R334W (1.2%) S945L (0.8%)
∆I507 (1.2%) K710X (0.8%)
2184delA (1.2%) 1078delT (0.8%)
R1158X (1.2%) Y122X (0.8%)
581
(Continued)
582
TABLE 1. Continued.
Estimated
BOBADILLA ET AL.
Region Mutation array detectiona mutationsb includedc (max/min)d Reference
Germany ∆F508 (71.8%) 1789+5G→A (0.9%) 87.6 76.7 17 5662/1316 Dörk et al. [1992]; Dörk et al.
R553X (2.0%) 3272-26A→G (0.9%) [1994]; Tümmler et al. [1996];
N1303K (1.8%) W1282X (0.7%) Estivill et al. [1997]; Dörk et
G542X (1.2%) 2143delT (0.7%) al. [2000]
R347P (1.2%) 1078delT (0.6%)
CFTRdele2,3 (1.2%) 2183AA→G (0.6%)
3849+10KbC→T (1.0%) 2184insA (0.6%)
G551D (0.9% 3659delC (0.6%)
1717-1G→A (0.9%)
Greece ∆F508 (52.9%) 3272-26A→G (0.8%) 82.2 67.6 22 2097/718 Kanavakis et al. [1995]; Estivill
621+1G→T (5.0%) R1070Q (0.8%) et al. [1997]; Tzetis et al.
G542X (4.1%) W496X (0.7%) [1997]; Macek et al. [2002]
N1303K (3.3%) 621+3A→G (0.7%)
2183AA→G (1.8%) ∆I507 (0.7%)
2789+5G→A (1.7%) W1282X (0.7%)
E822X (1.6%) 574delA (0.7%)
R117H (1.2%) 1677delTA (0.7%)
R334W (1.1%) A46D (0.6%)
R1158X (1.0%) 3120+1G→A (0.6%)
G85E (1.0%) G551D (0.5%)
Hungary ∆F508 (54.9%) W1282X (1.8%) 68.3 46.6 9 1133/976 CFGAC [1994]; Estivill et al.
1717-1G→A (1.9%) G542X (1.7%) [1997]; Macek et al. [2002]
R553X (2.1%) N1303K (1.3%)
Y1092X (1.8%) G551D (1.0%)
S1196X (1.8%)
Ireland ∆F508 (70.4%) G542X (1.0%) 82.1 67.4 7 801/509 CFGAC [1994]; Estivill et al.
G551D (5.7%) 621+1G→T (0.8%) [1994]
R117H (2.4%) 1717-1G→A (0.6%)
R560T (1.2%)
Italy ∆F508 (50.9%) ∆I507 (0.65%) 60.3 36.4 9 3524 Estivill et al. [1997]
(total) G542X (3.1%) W1282X (0.62%)
1717-1G→A (1.6%) Y122K (0.59%)
N1303K (1.4%) G551D (0.53%)
R553X (0.94%)
Italy ∆F508 (47.6%) R553X (1.3%) 87.1 75.9 15 225 Bonizzato et al. [1995]
(Northeast) R1162X (9.8%) 2789+G→A (1.3%)
2183AA→G (9.3%) Q552X (1.3%)
N1303K (4.0%) 621+1G→T (0.9%)
G542X (2.7%) W1282X (0.9%)
711+5G→A (2.7%) 3132delTG (0.9%)
1717-1G→A (2.2%) 2790-2A→G (0.9%)
G85E (1.3%)
Italy ∆F508 (56.4%) 711+1G→T (1.3%) 85.7 73.4 13 660/396 Castaldo et al. [1996]; Castaldo
(southern) N1303K (6.8%) G1244E (1.3%) et al. [1999]
G542X (5.7%) R1185X (1.3%)
W1282X (3.8%) L1065P (1.3%)
1717-1G→A (2.3%) R553X (1.1%)
2183AA→G (1.9%) I148T (0.7%)
4016insT (1.8%)
Latvia 1) DF508 (58.3%) 4) CFTRdele2,3 (2.8%) — — 6 36 Dörk et al. [2000]; Macek et al.
2) 3849+10KbC®T (8.3%) 5) W1282X (2.8%) [2002]
3) N1303K (5.6%) 6) 394delTT (2.8%)
Lithuania ∆F508 (31.0%) N1303K (2.0%) 39.0 15.2 4 94 Dörk et al. [2000]; Macek et al.
R553X (4.0%) CFTRdele2,3 (2.0%) [2002]
Macedonia ∆F508 (54.3%) 711+3A→G (1.0%) 69.2 47.9 12 559/226 Petreska et al. [1998]; Dörk et
G542X (4.2%) 3849G→A (1.0%) al. [2000]; Macek et al.
N1303K (2.0%) 2184insA (0.9%) [2002]
CFTRdele2,3 (1.3%) 457TAT→G (0.7%)
621+1G→T (1.3%) V139E (0.7%)
611-1G→T (1.2%) 1811+1G→C (0.6%)
Netherlands ∆F508 (74.2%) R1162X (0.9%) 86.8 75.3 9 3167/1442 Gan et al. [1995]; Estiville et al.
A455E (4.7%) S1251N (0.9%) [1997]; Collee et al. [1998]
G542X (1.8%) N1303K (0.9%)
1717-1G→A (1.5%) W1282X (0.7%)
R553X (1.2%)
Norway ∆F508 (60.2%) G551D (1.2%) 69.8 48.7 6 410/242 Schwartz et al. [1994]; Estivill
394delTT (4.2%) G542X (0.6%) et al. [1997]
R117H (3.0%) N1303K (0.6%)
Poland ∆F508 (57.1%) CFTRdele2,3 (1.8%) 73.5 54.0 11 4046/1726 CFGAC [1994]; Estivill et al.
3849+10Kb C→T (2.7%) R560T (1.5%) [1997]; Dörk et al [2000];
G542X (2.6%) W1282X (0.7%) Macek et al. [2002]
1717-1G→A (2.4%) ∆I507 (0.5%)
R553X (1.9%) G551D (0.5%)
N1303K (1.8%)
Portugal ∆F508 (44.7%) R334W (0.7%) 49.7 24.7 5 739/454 CFGAC [1994]; Estivill et al.
G542X (1.6%) N1303K (0.7%) [1997]
R1066C (2.0%)
Romania ∆F508 (36.6%) G542X (1.4%) 51.5 26.5 11 224/74 CFGAC [1994]; Estivill et al.
CFTR MUTATIONS
2043delG (2.0%) R553X (1.4%) [1997]; Popa et al. [1997];
W1282X (1.7%) G576X (1.4%) Macek et al. [2002]
1717-2A→G (1.4%) 1898+1G→A (1.4%)
I148T (1.4%) 2183AA→G (1.4%)
621+1G→T (1.4%)
Russia ∆F508 (54.4%) 552insA (0.9%) 70.7 50.0 12 5073/2562 CFGAC [1994]; Estivill et al.
CFTRdele2,3 (5.0%) G542X (0.9%) [1997]; Dörk et al. [2000];
R553X (3.5%) R334W (0.9%) Macek et al. [2002]
2183AA→G (1.3%) 1677delTA (0.8%)
W1282X (1.0%) Y122X (0.5%)
394delTT (1.0%) 1367del5 (0.5%)
583
(Continued)
584
TABLE 1. Continued.
Estimated
BOBADILLA ET AL.
Region Mutation array detectiona mutationsb includedc (max/min)d Reference
Slovakia ∆F508 (57.3%) CFTRdele2,3 (1.2%) 82.7 68.4 14 908/254 CFGAC [1994]; Estivill et al.
G542X (6.8%) 3849+10KbC→T (1.0%) [1997]; Dörk et al. [2000];
R553X (4.0%) S42F (0.9%) Macek et al. [2002]
N1303K (3.4%) R75X (0.9%)
2143delT (1.8%) G85E (0.9%)
R347P (1.4%) 605insT (0.9%)
W1282X (1.3%) 1898+1G→A (0.9%)
Slovenia ∆F508 (57.8%) R347P (1.1%) 79.7 63.5 16 455/132 CFGAC [1994]; Dörk et al.
2789+5G→A (4.1%) S4X (0.8%) [2000]; Macek et al. [2002]
R1162X (3.2%) 457TAT→G (0.8%)
G542X (1.9%) D192G (0.8%)
Q552X (1.5%) R553X (0.8%)
Q685X (1.5%) A559T (0.8%)
3905insT (1.5%) 2907delTT (0.8%)
CFTRdele2,3 (1.5%) 3667ins4 (0.8%)
Spain ∆F508 (52.7%) G85E (0.8%) 80.2 64.3 21 3608/1356 Chillón et al. [1994]; Casals et
G542X (8.0%) R1066C (0.8%) al. [1997]; Estivill et al. [1997]
N1303K (2.5%) 2789+5G→A (0.7%)
3601-111G→C (2.0%) 2869insG (0.7%)
1811+1.6Kb A→G (1.7%) ∆I507 (0.6%)
R1162X (1.6%) W1282X (0.6%)
711+1G→T (1.3%) L206W (0.5%)
R334W (1.2%) R709X (0.5%)
Q890X (1.0%) K710X (0.5%)
1609delCA (1.0%) 3272-26A→G (0.5%)
712-1G→T (1.0%)
Sweden ∆F508 (66.6%) E60X (0.6%) 85.9 73.8 10 1357/662 Schwartz et al. [1994]; Estivill et
394delTT (7.3%) Y109C (0.6%) al. [1997]; Schaedel et al.
3659delC (5.4%) R117H (0.6%) [1999]
175insT (2.4%) R117C (0.6%)
T338I (1.2%) G542X (0.6%)
Switzerland ∆F508 (57.2%) K1200E (2.1%) 91.3 83.4 9 1268/1173 Estivill et al. [1997];
R553X (14.0%) N1303K (1.2%) Hergersberg et al. [1997]
3905insT (9.8%) W1282X (1.1%)
1717-1G→A (2.7%) R347P (0.6%)
G542X (2.6%)
Ukraine ∆F508 (65.2%) CFTRdele2,3 (1.1%) 74.6 55.7 6 1055/580 Estivill et al. [1997]; Dörk et al.
R553X (3.6%) G551D (1.8%) [2000]; Macek et al. [2002]
N1303K (2.4%) W1282X (0.5%)
United ∆F508 (75.3%) 621+1G→T (0.93%) 81.6 66.6 5 19622/9815 Schwartz et al. [1995b];
Kingdom G551D (3.1%) 1717-1G→A (0.57%) Estivill et al. [1997]
(total) G542X (1.7%)
United ∆F508 (56.6%) 621+1G→T (1.8%) 69.1 47.7 7 456 CFGAC [1994]
Kingdom G551D (3.7%) R117H (1.5%)
(N. Ireland) R560T (2.6%) ∆I507 (0.9%)
G542X (2.0%)
United ∆F508 (19.2%) 621+2T→C (3.8%) 84.4 71.2 11 52 Malone et al. [1998]
Kingdom Y569D (15.4%) 2184insA (3.8%)
(Pakistani) Q98X (11.5%) R560S (1.9%)
1525-1G→A (9.6%) 1898+1G→T (1.9%)
296+12T→C (7.7%) R709X (1.9%)
1161delC (7.7%)
United ∆F508 (71.3%) 1717-1G→A (1.0%) 86.4 74.6 9 1236/730 Shrimpton et al. [1991];
Kingdom G551D (5.5%) 621+1G→T (0.6%) Gilfillan et al. [1998]
(Scotland) G542X (4.0%) ∆I507 (0.6%)
R117H (1.4%) R560T (0.6%)
P67L (1.4%)
United ∆F508 (71.6%) 1717-1G→A (1.1%) 98.7 97.4 17 183 Cheadle et al. [1993]
Kingdom 621+1G→T (6.6%) 3659delC (0.5%)
(Wales) 1898+1G→A (5.5%) R117H (0.5%)
G542X (2.2%) N1303K (0.5%)
G551D (2.2%) E60X (0.5%)
1078delT (2.2%) S549N (0.5%)
R1283M (1.6%) 3849+10KbC→T (0.5%)
R553X (1.1%) 4016insT (0.5%)
∆I507 (1.1%)
Yugoslavia ∆F508 (68.9%) 3849G→A (1.0%) 82.2 67.6 11 709/398 Dabovic et al. [1992]; Estivill et
G542X (4.0%) N1303K (0.8%) al. [1997]; Macek et al.
R1162C (3.0%) 525delT (0.5%) (submitted for publication)
457TAT→G (1.0%) 621+1G→T (0.5%)
I148T (1.0%) G551D (0.5%)
Q552X (1.0%)
Middle East/Africa
Algeria 1) DF508 (20.0%) 4) 1812-1G®A (5.0%) — — 5 20 Loumi et al. [1999]
2) N1303K (20.0%) 5) V754M (5.0%)
3) 711+1G®T (10.0%)
Jewish W1282X (48.0%) 3849+10KbC→T (6.0%) 95.0 90.3 6 261 Kerem et al. [1995]
(Ashkenazi) ∆F508 (28.0%) N1303K (3.0%)
G542X (9.0%) 1717-1G→A (1.0%)
CFTR MUTATIONS
Jewish 1) N1303K — — 1 6 Kerem et al. [1995]
(Egypt)
Jewish 1) Q359K/T360K — — 1 8 Kerem et al. [1995]
(Georgia)
Jewish 1) DF508 2) 405+1G®A — — 2 11 Kerem et al. [1995]
(Libya)
Jewish 1) DF508 (72.0%) 3) D1152H (6.0%) — — 3 33 Kerem et al. [1995]
(Morocco) 2) S549R (6.0%)
Jewish ∆F508 (35.0%) W1282X (2.0%) 43.0 18.5 4 51 Shoshani et al. [1992]
(Sepharadim) G542X (4.0%) S549I (2.0%)
585
(Continued)
586
TABLE 1. Continued.
Estimated
BOBADILLA ET AL.
Jewish 1) 405+1G®A (48.0%) 3) W1282X (17.0%) — — 4 23 Kerem et al. [1995]
(Tunisia) 2) DF508 (31.0%) 4) 3849+10KbC®T (4.0%)
Jewish 1) G85E 4) G542X — — 6 10 Kerem et al. [1995]
(Turkey) 2) DF508 5) 3849+10KbC®T
3) W1282X 6) W1089X
Jewish (Yemen) None — — 0 5 Kerem et al. [1995]
Lebanon 1) DF508 (35.0%) 6) 4096-28G®A (2.5%) — — 9 40 Desgeorges et al. [1997]
2) W1282X (20.0%) 7) 2789+5G®A (2.5%)
3) 4010del4 (10.0%) 8) M952I (2.5%)
4) N1303K (10.0%) 9) E672del (2.5%)
5) S4X (5.0%)
Reunion ∆F508 (52.0%) 1717-1G→A (0.7%) 90.4 81.7 9 138 Cartault et al. [1996]
Island Y122X (24.0%) G542X (0.7%)
3120+1G→A (8.0%) A309G (0.7%)
A455E (2.2%) 2789+5G→A (0.7%)
G551D (1.4%)
Saudi North: 3) H139L — — North 1 49 families El-Harith et al. [1997];
Arabia 1) 1548delG 4) L1177X Central 3 Kambouris et al. [1997];
Central: 5) DF508 South 4 Banjar et al. [1999]
1)I1234V 6) 3120+1G®A West 9
2)1548delG 7) 425del42 East 6
3)DF508 8) R553X
South: 9) N1303K
1) I1234V East:
2) 1548delG 1) 3120+1G®A
3) 711+1G®T 2) H139L
4) 3120+1G®A 3) 1548delG
West: 4) DF508
1) I1234V 5) S549R
2) G115X 6) N1303K
Tunisia ∆F508 (17.6%) G85E (2.6%) 58.7 34.5 11 78 Messaoud et al. [1996]
G542X (8.9%) W1282X (2.6%)
711+1G→T (7.7%) Y122X (1.3%)
N1303K (6.4%) T665S (1.3%)
2766del8NT (6.4%) R47W+D1270N (1.3%)
R1066C (2.6%)
Turkeye ∆F508 (24.5%) 1066L (1.3%) 80.6 65.0 36 1067/670 Yilmaz et al. [1995]; Estivill et al.
1677delTA (4.1%) E822X (1.3%) [1997]; Onay et al. [1998];
2789+5G→A (3.9%) 2183+5G→A+2184insA (1.3%) Macek et al. [2002]
2181delA (3.8%) D110H (0.8%)
R347H (3.6%) P1013L (0.8%)
N1303K (2.9%) 3172delAC (0.8%)
621+1G→T (2.6%) 1259insA (0.8%)
G542X (2.6%) M1028I (0.8%)
E92K (2.6%) 4005+1G→A (0.7%)
A96E (2.6%) W1282X (0.7%)
M152V (2.6%) I148T (0.6%)
2183AA→G (2.5%) R1162X (0.6%)
296+9A→T (1.6%) D1152H (0.6%)
2043delG (1.4%) W1098X (0.6%)
E92X (1.4%) E831X (0.6%)
K68N (1.4%) W496X (0.6%)
G85E (1.3%) F1052V (0.5%)
R1158X (1.3%) L571S (0.5%)
United Arab S549R (61.5%) ∆F508 (26.9%) 88.4 78.1 2 86/52 Frossard et al. [1988];
Emirates Frossard et al. [1999]
North/Central/South Americas
Argentina ∆F508 (58.6%) N1303K (1.8%) 69.1 47.7 5 326/228 CFGAC [1994]; Chertkoff et al.
W1282X (3.9%) 1717-1G→A (0.9%) [1997]
G542X (3.9%)
Brazilf ∆F508 (47.7%) W1282X (1.3%) 66.8 44.6 10 820/500 CFGAC [1994]; Cabello et al.
(total) G542X (7.2%) G85E (1.3%) [1999]; Raskin et al. [1999];
R1162X (2.5%) R553X (0.7%) Bernardino et al. [2000]
R334W (2.5%) L206W (0.6%)
N1303K (2.4%) 2347delG (0.6%)
South East: >∆F508, G542X
South: >N1303K
Brazil ∆F508 (31.7%) N1303K (2.5%) 42.5 18.1 3 120 Parizotto and Bertuzzo [1997]
(Sao Paulo) G542X (8.3%)
Canada ∆F508 (59.0%) G542X (0.5%) 98.5 97.0 13 381/200 Rozen et al. [1992];
(Lac St. Jean) 621+1G→T (24.3%) N1303K (0.5%) De Braekeleer et al. [1998]
A445E (8.2%) Q890X (0.5%)
Y1092X (1.2%) S489X (0.5)
711+1G→T (1.0%) R117C (0.5%)
I148T (1.0%) R1158 (0.5%)
G85E (0.8%)
Canada ∆F508 (71.4%) ∆I507 (1.3%) 90.9 82.6 7 77 Rozen et al. [1992]
(Quebec City) 711+1G→T (9.1%) Y1092X (1.3%)
621+1G→T (5.2%) N1303K (1.3%)
A455E (1.3%)
CFTR MUTATIONS
Canada ∆F508 (70.9%) W1282X (0.9%) 82.0 67.2 10 632 Kristidis et al. [1992]
(Toronto) G551D (3.1%) R117H (0.9%)
G542X (2.2%) 1717-1G→A (0.6%)
621+1G→T (1.3%) R560T (0.6%)
N1303K (0.9%) ∆I507 (0.6%)
Chile ∆F508 (29.2%) R553X (4.2%) 33.4 11.2 2 72 Rios et al. [1994]
Columbia 1) DF508 (35.4%) 3) N1303K (2.1%) — — 4 48 Restrepo et al. [2000]
2) G542X (6.3%) 4) W1282X (2.1%)
Ecuador 1) DF508 (25%) — — 1 20 Paz-y-Mino et al. [1999]
587
(Continued)
588
TABLE 1. Continued.
Estimated
BOBADILLA ET AL.
Mexico ∆F508 (41.6%) G551S (0.5%) 75.5 57.0 35 374/194 Orozco et al.[1993]; Villalobos-
G542X (5.6%) 1078delT (0.5%) Torres et al. [1997]; Liang et al.
∆I507 (2.5%) Y1092X (0.5%) [1998]; Orozco et al. [2000]
S549N (1.9%) R117H (0.5%)
N1303K (1.7%) G85E (0.5%)
R75X (1.5%) 1716G→A (0.5%)
406-1G→A (1.5%) W1204X (0.5%)
I148T (1.5%) W1098C (0.5%)
3849+10KbC→T (1.5%) 846delT (0.5%)
621+1G→T (1.2%) P750L (0.5%)
2055del9→A (1.0%) V754M (0.5%)
935delA (1.0%) R75Q (0.5%)
I506T (1.0) W1096X (0.5%)
3199del6 (1.0%) L558S (0.5%)
2183AA→G (1.0%) 4160insGGGG (0.5%)
G551D (0.5%) 297-1G→A (0.5%)
R553X (0.5%) H199Y (0.5%)
1924del7 (0.5%)
United States ∆F508 (68.6%) R553X (0.9%) 79.7 63.5 10 25048 Cystic Fibrosis Foundation
(total) G542X (2.4%) 621+1G→T (0.9%) [1998]
G551D (2.1%) 1717-1G→A (0.7%)
W1282X (1.4%) 3849+10KbC→T (0.7%)
N1303K (1.3%) R117H (0.7%)
United States ∆F508 (48.0%) S1255X (1.4%) 77.3 59.8 16 160/148 Carles et al. [1996]; Macek et al.
(African 3120+1G→A (12.2%) 444delA (0.7%) [1997]; Dörk et al. [1998];
American) 2307insA (2.0%) R334W (0.7%) Friedman et al. [1998]
A559T (2.0%) ∆I507 (0.7%)
R553X (2.0%) 1717-1G→A (0.7%)
∆F311 (2.0%) G542X (0.7%)
G480C (1.4%) S549N (0.7%)
405+3A→C (1.4%) G551D (0.7%)
United States 1) L1093P — — 1 2 Yee et al. [2000]
(Cherokee)
United States Non-French: French: Non- Non- Non- Non- Bayleran et al. [1996]
(Maine) ∆F508 (82.0%) ∆F508 (58%) French: French: French: French:
G542X (2.6%) 711+1G→T (8.3%) 95.3 90.8 11 191
G551D (2.6%) I148T (4.2%) French: French: French: French:
N1303K (2.1%) A455E (4.2%) 80.3 64.5 8 72
R560T (1.0%) 1717-1G→A (1.4%) Total:
621+1G→T (1.0%) G85E (1.4%) 263
711+1G→T (1.0%) 621+1G→T (1.4%)
R117H (1.0%) Y1092X (1.4%)
1717-1G→A (1.0%)
G85E (0.5%)
W1282X (0.5%)
United States ∆F508 (46.0%) R334W (1.6%) 58.5 34.2 7 129 Grebe et al. [1994]
(SW Hispanic) G542X (5.4%) W1282X (0.8%)
3849+10KbC→T (2.3%) R553X (0.8%)
R1162X (1.6%)
United States 1) R1162X — — 3 17 Mercier et al. [1992]
(SW Native 2) D648V
American) 3) G542X
United States 1) R1162X 3) G542X — — 4 16 Mercier et al. [1994]
(Zuni Pueblo) 2) 3849+10KbC®T 4) D648V
Venezuela ∆F508 (29.6%) G542X (3.7%) 33.3 11.1 2 54 Restrepo et al. [2000]
Other Regions
Australia ∆F508 (76.9%) 621+1G→T (1.1%) 88.7 78.7 8 761/464 CFGAC [1994]
G551D (4.5%) N1303K (0.9%)
G542X (2.8%) W1282X (0.6%)
R553X (1.3%) R117H (0.6%)
East Asia 1) 1898+1G®T 2) 1898+5G®T — — 2 28 Suwanjutha et al. [1998]
Hutterite 1) M1101K (69.0%) 2) DF508 (31.0%) — — 2 32 Zielenski et al. [1993]
Brethren
New Zealand ∆F508 (78.0%) N1303K (1.9%) 87.4 76.4 5 636 CFGAC [1994]
G551D (4.4%) 621+1G→T (1.1%)
G542X (2.0%)
*This table presents the mutation panels for all regions investigated in this study. Countries and regions are sorted by geographic location and appear in alphabetical order within their
geographical divisions. Overall mutation detection rates, as well as theoretical projections of screening diagnosis rates, appear in columns 3 and 4. Those regions with an nc<50 appear in italics.
It should be noted that the R117H mutation is only associated with classical CF when associated with the 5T variation. Therefore, identification of the R117H mutation should prompt investigation
for the 5T variant as well, R117H (IVS8-5T), and this combination should be regarded as a single CF causing allele.
a
This column presents the total percent of CF chromosomes that have been completely genotyped with identifiable mutations.
b
Using data from column 3, the estimated CF diagnosis based solely on genetic analysis was calculated. Assuming Hardy-Weinberg equilibrium.
c
Listed here are the total number of mutations that account for the percentages reported in columns 3 and 4.
d
This column gives the number of chromosomes that the arrays and detection/diagnosis rates are based on. They are presented in a maximum/minimum format. In those regions where only one
study was performed, only the absolute chromosome number appears, as there is no possibility of sample overlap.
e
Onay et al. [2001] confirm the aggregated database calculated from previous publications, revealing nearly identical percentages and strengthening conclusions regarding CFTR heterogeneity
in Turkey.
f
Cabello et al. [2001] also reported 3120+G→A in five patients with African ancestry living in Rio de Janeiro.
CFTR MUTATIONS
589
3120+1G→A. This correlates to the 27% Eu- 25.0 (95% CI, 17.732.3) mutations account for
ropean and 40% African origins of the original 84.0% (95% CI, 77.990.1) of the CF alleles in
founders of this isolated island [Cartault et al., these regions. This is most likely due to their
1996]. geographic situation, serving as historic gate-
A theoretical maximum number of CF chro- ways into Europe from the Middle East, Africa,
mosomes analyzed in this study is 119,257, as- and associated waterways.
suming 0% overlap in regions with multiple We performed a comparative analysis of these
studies where overlap could not be definitively four European countries (Turkey, Bulgaria,
excluded. The absolute minimum total number Greece, and Spain) that are situated with coastal
of distinct CF chromosomes reviewed in this territories that enabled them to function as
study consisted of 72,431; this number is derived gateways to the continent over long periods of
assuming 100% chromosome overlap in regions ancient Neolithic migration. This revealed a sig-
where multiple studies have been conducted. nificantly higher total number of mutations
Table 1 presents a listing of all regions that present in the gateway countries when evalu-
had CFTR genotype information available. In ated against the remaining European countries.
those regions with a low number of chromo- When comparing the number of mutations
somes, less than 50 (nc<50), the results appear present in this group of gateway countries ver-
in italics. These data were compiled and the rela- sus the others (25.0 mutations vs. 10.2 muta-
tive number of CFTR alleles detected were de- tions), a statistically significant difference was
termined on a regional basis. The detection rates observed with a p-value = 0.001; however, the
ranged from a maximum of 100% in Belgium and total percentage of CFTR mutations identified
98.7% in Brittany, France, and Wales, to a mini- is not significantly different (p = 0.48). While it
mum of 33.3% in Venezuela. When these results may be possible to account for the 5.1% detec-
are ranked, seven regions had detection rates of tion discrepancy by evoking a theory of multiple
95% or greater, using mutations present at mutations present at a very low frequency (on
greater than 0.5% in the CF patient pool. By average, 15 mutations at ∼0.34% each), it seems
decreasing the threshold to 90%, we found that unlikely that this is the case due to the vast varia-
six more regions could be included, for a total of tion in mutation array size (2.5 times larger in
13. In addition, 21 more regions were identified the gateway regions) and the exceedingly well
with detection rates above 80%. These results investigated genomes of the non-gateway na-
listed in Table 1 are estimates based solely on tions. This discrepancy is more likely explained
those mutations that have been identified at by the presence of a few unidentified mutations
0.5% or greater. Actual detection rates could be present at relatively higher frequency and/or
higher; however, in most cases, this would re- rounding artifact in the non-gateway regions. In
quire the screening of many more mutations all, these data tend to support the concept that
present at very low prevalence. these countries did indeed serve as gateways
into Europe. Over the centuries, CFTR muta-
European Spectrum
tions have fluxed into and out of Europe through
Central, northern, western, and northeastern these regions, and the effects of this migration
Europe (including Austria, Belarus, Belgium, can be seen in the high mutational heterogene-
Denmark, Estonia, Finland, France, Germany, ity of the current CF patient population. This
Lithuania, the Netherlands, Norway, Poland, diverse panel of CFTR mutations further com-
Russia, Sweden, Switzerland, and the Ukraine) plicates the ability of molecular tests to detect
show a large degree of homogeneity among CF patients confidently with any screening pro-
CFTR mutations (Table 1). On average, 10.2 gram. Therefore, these regions are prime candi-
(95% CI, 7.413.0) mutations per country ac- dates for the implementation of screening
count for 78.9% (95% CI, 72.185.7) of the to- protocols that can detect a wide range of unre-
tal number of CF chromosomes in these regions. lated mutations.
This is largely attributable to the high prevalence G542X is most common in the Mediterranean
of ∆F508 in these regions. Spain, Bulgaria, regions of Europe and Africa. Italy and France
Greece, and Turkey have some of the most di- (Table 1) show a gradient of diffusion that is seen
verse mutational arrays in Europe. On average, commonly across southern Europe. Regions such
as extreme northwest France (Brittany) have a minishing frequencies as one moves to the south-
greater than five-fold decrease in this mutation ern and the eastern portions of Europe. One ini-
frequency when compared to the Mediterranean tially puzzling phenomenon uncovered by CFTR
coastal part of southern France (Brittany, 1.1%; mutation analysis is the relatively high incidence
Southern France, 6.4%; p = 0.001). Spain also of this mutation in the Czech Republic (3.8%,
displays this trend (Mediterranean Coastal Table 1). In order to explain this variance, it is
Spain, 14.4%; all other regions of Spain, 5.7%; p necessary to analyze the migration of early Celtic
<0.001) [Casals et al., 1993]. This gradient of and Germanic tribes and the development of the
decreasing G542X prevalence can be seen on Roman Empire. In the Neolithic period, Celts
both the international and intranational levels once inhabited a large band of land spanning
as the distance from the Mediterranean Sea in- across much of western, central and eastern
creases. Some countries, however, like Hungary, Europe, just north of the Alps. During the last
Slovakia, and Austria, tend to show a more century BCE, both the Roman Empire and the
Mediterranean mutation panel. This can be par- Germanic tribes expanded and displaced or ab-
tially attributed to the retrograde colonization sorbed Celtic tribes. This expansion eventually
of many portions of Europe via waterways like resulted in the expulsion of the Celts into what
the Danube river [Serre et al., 1990]. Early colo- are now present-day Ireland, UK, and northwest
nization of Europe via the branching waterways France. However, a few tribes (e.g., the Boii)
carried these Mediterranean mutations up into maintained a relatively small region just north
the continent of Europe, where they now are of the Alps, and assimilated with incoming popu-
present at unexpectedly higher incidences. lations [James, 1993]. This region was eventu-
G542X has been implicated as a mutation that ally known as Bohemia, which now corresponds
was introduced into the Mediterranean region to the Czech Republic and some portions of
by the migration of Phoenicians [Loirat et al., southeast Germany and northern Austria
1997]. Across Europe, G542X is found in sig- [Macek et al., 1991]. The effects of this move-
nificantly higher proportions in countries, and ment can be seen today by the nearly three times
in regions of countries, which border the Medi- higher than expected frequency of G551D in the
terranean Sea (Table 1). Not surprisingly, G542X Czech Republic.
has its highest rates of prevalence in the south W1282X is a mutation of single origin that
of Spain and in the north of Africa (South Spain, has historically been associated with the
14.4%; Tunisia, 8.9%). Interestingly enough, Ashkenazi Jews. This mutations frequency has
these two regions correspond to two major an- been amplified to almost double that of the
cient Phoenician cities, Carthagène and Carthage, ∆F508 mutation in this distinct population
respectively [Loirat et al., 1997]. The genetic re- [Kerem et al., 1995]. As with all other popula-
sults of the introduction of this mutation into such tion specific mutations, the W1282X mutation
a busy trade and trafficking center as the Medi- is seen within the mutational arrays of the mul-
terranean can still be observed today. When one titude of countries that have had a significant
studies CF chromosomes in those populations Ashkenazi Jewish influence.
of the New World that were heavily influenced 394delTT, referred to as the Nordic muta-
by such countries as Spain and Italy, the G542X tion [Schwartz et al., 1994], has been found to
mutation frequencies are virtually identical. The be present at a high frequency in the countries
mutation N1303K is another CFTR allele that bordering the Baltic Sea and associated water-
has a similar distribution patterns as G542X, and ways (Sweden, Norway, Denmark, Finland, Es-
may also have been introduced via the Mediter- tonia, Russia, etc.). 394delTT is associated
ranean route (Table 1). almost exclusively with a single chromosomal
Another mutation that seems to have a clear haplotype, which suggests a single origin, cen-
history is the G551D allele. This allele has char- tered in this region [Schwartz et al., 1994]. Many
acteristically been associated with populations of the present day populations in this region have
that are of a Celtic descent, and is seen at its a common ancestry, and therefore, this muta-
highest prevalence in regions such as Ireland and tion is rare outside these populations.
Brittany. This mutation has also diffused across Another CFTR mutation that is common in
the remainder of Europe, and can be seen in di- central and eastern Europe is the CFTR dele2,3
(21kb) [Dörk et al., 2000]. To date it has been patients. The top 10 list for the United States
found in many regions of Europe with a strong (Table 1) clearly does not include this muta-
Eastern Slavic origin, including the Czech Re- tion, further stressing the importance of com-
public, Russia, Belarus, Austria, Germany, Po- plete characterization of a population prior to
land, Slovenia, Ukraine, and Slovakia. It has also implementing multi-mutation screening pro-
been found in a few other regions of Europe, but grams. We include 3120+1G→A in our rec-
is absent in the vast majority of the remaining ommended panel of mutations for the United
countries. States in 2001 (Table 2).
Eastern Canada shows a wide variability in
North American Spectrum
CF-causing mutations spread over a relatively
Table 1 also shows the data from North small geographic region (Table 1). With an ap-
America. In the United States, the CFTR alle- parent average incidence of 1 per 914 births, the
les reflect the geographic origin of the current Lac St. Jean (LSJ) region of Quebec Canada has
population with a strong relationship to Euro- the highest CF incidence of all the populations
peans. The results from analysis of 25,048 chro- assessed in this study [Rozen et al., 1992; De
mosomes indicate that there are at least 10 Braekeleer et al., 1998]. LSJ also presents with a
CFTR alleles found at greater than 0.5% fre- relatively low ∆F508 level (59%), a high
quency in CF chromosomes of patients resid- 621+1G→T prevalence (24.3%), and the sec-
ing in the United States. These 10 mutations ond highest CF allele detection rate identified
account for 79.7% of the identified CFTR mu- in this study (98.5%). Historically, the LSJ re-
tations. The top 10 list for the United States gion of Canada was isolated geographically, and
(Table 1) includes five CFTR alleles found in had a small founder population [Rozen et al.,
populations with distinct ethnic ancestries, i.e., 1992]. These factors most likely account for the
G542X, G551D, W1282X, N1303K, and large deviation noted when compared to other
3849+10KbC→T. nearby Canadian cities.
The CFTR mutation 3120+1G→A is the When compared to Toronto and Quebec City,
second most prevalent allele in African Ameri- LSJ has a statistically lower ∆F508 percentage
can populations of CF patients and is second (p =0.002 for Toronto vs. LSJ; p = 0.047 for
only to the ∆F508 mutation, which presumably Quebec vs. LSJ; Quebec vs. Toronto p = 0.93)
emerged via ethnic admixture with Caucasians and a statistically higher 621+1G→T percent-
(Table 1). This mutation has also been found age (p = 0.00001 for Toronto vs. LSJ; p =
in Native African and Arab populations [El- 0.00001 for Quebec vs. LSJ; Quebec vs. Toronto
Harith et al., 1997; Kambouris et al., 1997; p = 0.12). The total percentage of CF muta-
Banjar et al., 1999]. For instance, CF chromo- tions identified is also higher in LSJ compared
somes from black CF patients of South Africa to Toronto or Quebec City (98.5% vs. 82.0% (p
show an estimated frequency of 46%, making = 0.00001) and 90.9% (p = 0.026), respec-
3120+1G→A the most common mutation in tively). The investigation of Toronto utilized a
that population where ∆F508 was not found 25 mutation reverse dot blot panel, yielding a
[Goldman et al., 2001]. With only a few ex- total detection rate of 82.0% [Kristidis et al.,
ceptionsmainly regions that had early con- 1992]; and a study by Rozen et al. [1992] com-
tributions of African founder genesthis pared LSJ and Quebec with a 10 mutation panel.
mutation is rarely seen outside the African mu- Results of this study revealed mutation identifi-
tation array [Macek et al., 1997]. 3120+1G→A cation rates of 93% and 91%, respectively. Newer
is a prime example of the impact that a single techniques such as DDGE have been applied to
rare mutation can have on the efficiency of a the LSJ region recently [De Braekeleer et al.,
screening program. This mutation, which is tra- 1998], and this has allowed for the statistically
ditionally isolated to African descended gene significant increase in mutation detection rates
pools, will account for over 12% of the total identified in this study. LSJ serves a prime ex-
CFTR alleles in African American CF patients, ample of how thorough characterization of mu-
and overlooking it when screening newborns tational diversity can help increase overall
for CF could lead to lower sensitivity and po- population detection rates, and strengthen po-
tentially delayed diagnosis for many of these tential screening panels.
TABLE 2. Ideal Recommended CFTR Mutation Screening Panel for 2001 Neonatal Screening in the USA*
Location Estimated
Mutation in CFTRa percentageb Reason for inclusion
DF508 Exon 10 68.6% CFF registry, >1%, Pan-European
G542X Exon 11 2.4% CFF registry, >1%, Mediterranean
G551D Exon 11 2.1% CFF registry, >1%, Celtic
W1282X Exon 20 1.4% CFF registry, >1%, Ashkenazi Jew
N1303K Exon 21 1.3% CFF registry, >1%, Mediterranean
R553X Exon 11 0.9% CFF registry, >0.5%, Hispanic
621+1G®T Intron 4 0.9% CFF registry, >0.5%, multi-ethnic
1717-1G®A Intron 10 0.7% CFF registry, >0.5%, Italian
3849+10KbC®T Intron 19 0.7% CFF registry, >0.5%, Hispanic
R117Hc Exon 4 0.7% CFF registry, >0.5%
1898+1G→T Intron 12 0.4% CFF registry, >0.1%, East Asian
DI507 Exon 10 0.3% CFF registry, >0.1%, Hispanic
2789+5G®A Intron 14b 0.3% CFF registry, >0.1%
G85E Exon 3 0.3% CFF registry, >0.1%
R347P Exon 7 0.2% CFF registry, >0.1%
R334W Exon 7 0.2% CFF registry, >0.1%, multi-ethnic
R1162X Exon 19 0.2% CFF registry, >0.1%, multi-ethnic
R560T Exon 11 0.2% CFF registry, >0.1%
3659delC Exon 19 0.2% CFF registry, >0.1%
A455E Exon 9 0.2% CFF registry, >0.1%
2184delA Exon 13 0.1% CFF registry, >0.1%
S549N Exon 11 0.1% CFF registry, >0.1%, multi-ethnic
711+1G®T Intron 5 0.1% CFF registry, >0.1%
R75X Exon 3 0.2% Hispanic
406-1G→A Intron 3 0.2% Hispanic
I148T Exon 4 0.2% Hispanic, French
2055del9→A Exon 13 0.1% Hispanic
935delA Exon 6b 0.1% Hispanic
I506T Exon 10 0.1% Hispanic
3199del6 Exon 17a 0.1% Hispanic
2183AA→G Exon 13 0.1% Hispanic
3120+1G®A Intron 16 1.5% African American, Arabian
2307insA Exon 13 0.2% African American
A559T Exon 11 0.2% African American
∆F311 Exon 7 0.2% African American
G480C Exon 10 0.2% African American
405+3A→C Intron 3 0.2% African American
S1255X Exon 20 0.2% African American
L1093P Exon 17b Undetermined Native American
D648V Exon 13 Undetermined Native American
I1234V Exon 19 Undetermined Arabian linkage
S549R Exon 11 Undetermined Arabian linkage
1898+5G→T Intron 12 Undetermined East Asian linkage
CFTRdele2,3 Exons 2,3 Undetermined Eastern European linkage (Slavic)
Y1092X Exon 17b Undetermined French linkage
394delTT Exon 3 Undetermined Nordic linkage
Y569D Exon 12 Undetermined Pakistani linkage
3905insT Exon 20 Undetermined Swiss linkage (also: Amish, Acadian, Mennonite)
1898+1G®A Intron 12 Undetermined Welsh linkage
M1101k Exon 17b Undetermined Hutterite ancestry
*This table presents the top 50 mutations in the USA based on the Cystic Fibrosis Foundation CF Registry data from 1997
[Cystic Fibrosis Foundation, 1998], and data generated during our investigation. Mutations that are highly prevalent in certain
ethnic populations are included to help boost the overall detection rate and help guard against internal screening biases. Those
mutations that are bolded are included in the ACOG/ACMG recommendations; note that 1078delT was excluded from the list
above because of its low frequency in the USA [Cystic Fibrosis Foundation, 1998].
a
This column indicates the location of the mutation within the introns and exons of the CFTR gene.
b
This column presents the estimated contribution of the individual mutations to the total US CF mutation pool. Those mutations
listed from the CFF Registry are absolute percentages based on the 1997 CF population. For those mutations assigned to a
particular ethnic population, values were estimated based on the prevalence of the mutation within the group and the fraction
that that group contributes to the total US population, as listed by the 2000 census [US Census Bureau, 2000b].
c
The mutation R117H is only associated with classical CF when associated with the 5T variation; therefore, identification of the
R117H mutation should prompt investigation for the 5T variant as well, R117H(IVS8-5T).
Central and South American Spectrum can account for at least half the CF patients (e.g.,
CFTR mutation characterization in the Cen- North America).
tral and South American countries to date has After all the regional data were gathered and
been quite sparse. Generally, there have been sorted, we performed scatter plot analysis with
very few studies that examined the CFTR mu- linear regression extrapolation. Filter 1 was used
tation spectrum in Central and South America. for the ∆F508 percentage vs. CF incidence analy-
In those studies that have assessed this region of sis and filters 1 and 2 were used for the remain-
the world, the results were limited, and a major- ing two analyses. Table 3 presents the filtered
ity of mutations went unidentified. During the data used in the regression analysis section of
course of this study, we were able to generate this study. The remaining regions are ranked by
enough data to examine CFTR mutational dis- incidence wherein the incidence is represented
tributions across South America, as shown in as one case per x live births. This table presents
Table 1. For comparative analysis, South the ∆F508 frequency, the total number of muta-
America was divided into mountainous and tions that are present at a frequency of 0.5% or
highland regions. By dividing South America greater, and the relative CF incidence for each
arbitrarily into two components, the west coast country or region included in the final analysis.
mountainous (Colombia, Ecuador, and Chile) When plotting the resultant data (Fig. 2), a sta-
region and east coast highlands (Venezuela, tistically significant relationship was found be-
Brazil, and Argentina) region, we found a sig- tween ∆F508 and CF incidence (p = 0.01, R2 =
nificant (p=0.00001) difference in the ∆F508 0.25). Thus, the variability in CF incidence
frequencies, namely 29.9% (nc=140 chromo- among countries correlates with the variability
somes) vs. 45.3% (nc=782 chromosomes), re- of the ∆F508 mutation. Other studies have il-
spectively. This variation may be attributable to lustrated this correlation as well [Lucotte et al.,
early colonization patterns of South America. 1995]. Although the strength of the relation-
The east coast highland nations of South ship may seem less than expected, it should be
America typically have a greater tie to European recognized that the limited precision of CF inci-
colonization, with major contributions from Por- dence data and the incompleteness of CFTR
tuguese, Spanish, and Jewish peoples. Not sur- allelic profiles may account for the predictive
prisingly, the mutational arrays of the highland powers of ∆F508 alone. In addition, the inci-
region mimic those of the major southern Euro- dence of CF across countries is influenced by
pean exploratory nations. In those regions that numerous non-genetic factors, which themselves
are isolated by the Andes mountain ranges (Co- vary markedly on a worldwide basis (e.g., envi-
lumbia, Ecuador, and Chile), CFTR allele ronmental and clinical determinants such as
characterization is more difficult, despite several under diagnosis). There have been many specu-
efforts (see Table 1 references). This is presum- lations as to the reason for the high prevalence
ably due to the greater association with of ∆F508 and CF in Europe. One of the leading
Amerindian gene pools and their typical CFTR hypotheses is that of heterozygote advantage,
mutations. This hypothesis is further strength- stating that the carrier state of CF in some way
ened by the relative absence of the ∆F508 mu- gave a selective advantage to those who carried
tation from the mutations listed for the three a single flawed CFTR allele. Possible mechanisms
Native American tribes that have thus far been include some resistance to diarrheal illnesses like
studied (Table 1). cholera, which directly stimulate the CFTR
channel [Romeo et al., 1989].
REGRESSION ANALYSES
The apparent correlation between ∆F508 and
Scatter plot and regression analysis of ∆F508 CF incidence was tested using two well-charac-
versus incidence and number of mutations pro- terized CF populations that have been studied
vided interesting results, which may help to close over more than 15 years by newborn screening
a gap that has been not yet been fully resolved and genotyping in Brittany, France, and Wiscon-
in the literature. Currently, there has been some sin. These two regions were not used in gener-
dispute as to the impact of ∆F508 on the total ating the linear regression model, and thus were
CF mutational diversity and CF prevalence. It examined to assess the validity of the new model.
is obvious that in some regions the ∆F508 allele Using the regression model, Wisconsin was pro-
TABLE 3. Filtered Data for CF Incidence Across the Globe*

Incidence
Country or (1 case per Mutations ∆F508
population X births)a >0.5%b percentb Reference
These regions passed both Filter 1 and Filter 2:
Finland 25000 4 46.2 Lucotte et al. [1995]; Kere et al. [1994]
Mexico 8500 15 41.6 Grebe et al. [1994]
Sweden 7300 10 66.6 Lucotte et al. [1995]; Schaedel et al. [1999];
Romeo et al. [1989]
Poland 6000 11 57.1 Lucotte et al. [1995]
North Ireland 5350 7 56.6 Romeo et al. [1989]
Russia 4900 12 54.5 Lucotte et al. [1995]
Denmark 4700 6 87.5 Lucotte et al. [1995]
Norway 4500 6 60.2 Lucotte et al. [1995]
Netherlands 3650 9 74.2 Lucotte et al. [1995]; Collee et al. [1998]
Spain 3500 21 52.7 Lucotte et al. [1995]
Greece 3500 22 52.9 Lucotte et al. [1995]
Germany 3300 17 71.8 Lucotte et al. [1995]; Romeo et al. [1989]
United States 2835 10 68.6 Kosorok et al. [1996]; Romeo et al. [1989]
Czech Republic 2833 10 70.0 Lucotte et al. [1995]; Romeo et al. [1989]
United Kingdom 2600 5 75.3 Lucotte et al. [1995]; Romeo et al. [1989]
Australia 2500 8 76.9 Romeo et al. [1989]
Italy 2438 9 50.9 Lucotte et al. [1995]; Gasparini et al. [1999];
Romeo et al. [1989]
France 2350 12 67.7 Lucotte et al. [1995]
Switzerland 2000 9 57.2 Hergersberg et al. [1997]
Ireland 1800 7 70.4 Lucotte et al. [1995]
The following five regions did not pass Filter 2 Criterion:
United Arab Emirates 15876 n/a 26.9 Frossard et al. [1998]; Frossard et al. [1999]
Brazil 6902 n/a 47.7 Cabello et al. [1999]; Raskin et al. [1999];
Bernardino et al. [2000]
Estonia 4500 n/a 51.7 Klaassen et al. [1998]
Chile 4000 n/a 29.2 Rios et al. [1994]
Scotland 1984 n/a 71.3 Shrimpton et al. [1991]; Gilfillan et al. [1998]
*This table shows the final data set for CF incidences after filtering for incomplete data sets (filter 1) and mutation detection
methodology (filter 2). The results are ranked by incidence from lowest to highest. Incidence is represented as 1 Case per X
births (Column 2). Column 3 presents the data that was generated during phase one of this study. This column presents the total
number of mutations, including ∆F508, present in the listed population that exceed the 0.5% threshold. Those regions in the
lower section of the table did not have methods strong enough to pass Filter 2, and therefore were only used in the ∆F508 versus
Incidence regression analysis.
a
CF incidence values were derived from those references listed in column 5.
b
Data derived from this investigation, see Table 1 for applicable references.
jected to have a CF incidence of 1:2,976 [95% Using the data presented in Table 3 and Fig-
CI = 1:2,5001:3,676], given a ∆F508 preva- ure 2, we illustrate that ∆F508 plays a significant
lence of 71.25% [Gregg et al., 1997]. The actual role in determining at least one of these variables,
Caucasian CF incidence in Wisconsin is 1:3,419 CF incidence. By performing regression analysis,
[Kosorok et al., 1996]. Brittany was projected we were able to estimate that ∆F508 is contribut-
to have a CF incidence of 1:2,309 (95% CI = ing at least 1/4 of the total influence, on average,
1:2,3091:3,508) based on a ∆F508 prevalence to CF incidence. Regression analysis allowed for
of 75.8%. The actual incidence is 1:2,913 [Scotet the approximation of CF incidence based on
et al., 2000]. ∆F508 prevalence in two well-characterized popu-
There also appeared to be an inverse correla- lations. Clearly, this linear regression line will not
tion when examining the total number of muta- apply to all populations. Factors such as founder
tions greater than 0.5% with regard to ∆F508 effects, consanguineous unions, and random hu-
prevalence; however, this relationship was not man migration all tend to shift any gene or muta-
statistically significant (p = 0.15, R2 = 0.11). tion pool away from Hardy-Weinberg equilibrium.
There was no relationship observed between the A prime example is the unexpectedly low
number of mutations greater than 0.5% and the ∆F508%, high S549R%, and 88.4% detection rate
population CF incidence values (p = 0.61, R2 = in the United Arab Emirates, a region that has
0.01), i.e., a greater variety of CFTR mutations experienced all of these confounding factors to
does not appear to influence disease incidence. some extent.
FIGURE 2. ∆F508 vs. incidence with linear regression analysis. This graph illustrates the relationship of ∆F508 to the CF
incidence. Dashed lines represent the 95% CI for incidence when the regression is modeled and analyzed to predict
incidence based on ∆F508 values. Analysis shows a significant positive correlation between the ∆F508% and the total
population CF incidence. Statistical analysis indicates that ∆F508 contributes roughly 1/4 of the total population CF
incidence on average. Those countries that are bolded were not included in the second tear of regression analysis due to
failure to meet inclusion criterion for filter 2.
SPECULATION AND FUTURE ISSUES ture further haplotype analysis of these mutation-
When examining the mutational arrays of sharing chromosomes will help to strengthen this
populations within the United States, two mu- concept of non-related recurrent mutations. The
tations are of special interest, namely R1162X need to probe for the 3849+10KbC→T muta-
and 3849+10KbC→T. Recent studies [Mercier tion in screening programs is even more impor-
et al., 1992, 1994] show that R1162X has been tant than including other minor CFTR alleles.
found in certain Native American populations, Patients with this mutation can present with no
as well as the U.S. Hispanic CF patient pool. elevation of sweat chloride levels, but sino-
3849+10KbC→T has been found in both the pulmonary disease is still moderate to severe. A
Mexican [Villalobos-Torres et al., 1997; Liang et delay in such a diagnosis on the genetic level could
al., 1998; Orozco et al., 2000] and the Southwest result in significant lung involvement by the time
U.S. Hispanic [Grebe et al., 1994], as well as cer- the diagnosis is made clinically.
tain Native American CF patients [Mercier et al., One of the major European genetic contribu-
1994], but is not a common mutation in Europe tors to the New World Latin population was
(Table 1). Those regions of Europe that have also Spain. Since the 3849+10KbC→T mutation is
been shown to exhibit this mutation (i.e., Poland not high in the Spanish population, it seems
and Ashkenazi Jewish populations) do so most plausible that the indigenous people of the
likely as a result of mutational recurrence [Morral Americas are the source of introduction of this
et al., 1994], and not a direct founder effect. Fu- mutation into the Hispanic mutation pool of the
Americas. This theory is further strengthened cutoff of 0.1% for the CFF data was used be-
by the observation that this mutation is the sec- cause of the exceedingly large sample size, over
ond most prevalent mutation in the Zuni Pueblo 25,000 chromosomes. Because of this, we are
population [Mercier et al., 1994]. By extrapo- confident of the prevalence of these mutations,
lating this hypothesis to other populations, it is and offer them as options in multi-mutation
reasonable to assume that 3849+10KbC→T will screening programs for the United States.
likely be present at higher percentages in other CFTR panels employing these 50 alleles can
Native American populations, as well as in those also be applied to screening newborn populations
countries in Central and South America that of Western Europe, but genotype profiles of spe-
have close ties to Amerindian ancestry (Ecua- cific regions should be examined carefully be-
dor and Chile in this study). Currently, the un- fore implementation [Farrell, 2000]. When
derstanding of CF-causing mutations is very relative weighting is done for each mutation,
limited with regard to the Native American, based on ethnic prevalence in the United States,
Central, and South American populations; how- the results indicate that the panel recommended
ever, by examining the Mexican and U.S. His- in Table 2 is expected to detect 86.3% of the CF
panic mutational arrays, it may be possible to chromosomes of U.S. patients. This is nearly a
predict which mutations were originally indig- 4% increase in detection based on the CFF Reg-
enous to the Americas. istry data (82.5%), and a 2.5% increase when
compared to the ACMG/ACOG panel (83.8%).
RECOMMENDATIONS AND APPLICATIONS
These detection rates are based on the total CF
Aside from the major mutations above, there population in the United States. Clearly, this
are many other CFTR alleles that are present expanded panel will have more of an impact on
only in specific populations or at variable per- the regional and state levels than the national
centages in different populations, i.e., M1101K level. Geographical variation of persons of dif-
in the Hutterite Brethren [Zielenski et al., 1993]. fering ethnic backgrounds will largely alter the
These are the minor mutations that take on application and contribution of the extended
special significance when one is selecting CFTR panel in many regions.
alleles to test for in screening programs. Achiev- When the panels are re-examined with an
ing high sensitivity in prenatal and neonatal ethnic frame in mind, the detection rates differ
screening, as well as avoiding inadvertent dis- significantly. Based on the data compiled in this
crimination, requires knowledge of regional an- paper, for African American CF patients, the
cestry and/or CFTR arrays. Some countries have ACMG/ACOG panel will detect an estimated
only a few mutations that account for the vast three CFTR mutations accounting for a 62.2%
majority of CF patients, but other countries have chromosomal identification rate. Our extended
many CFTR alleles. panel will encompass nine mutations in this group
Based on our comprehensive assessment, we and extend the detection rate to 72.4% (a 10.2%
offer recommendations (Table 2) for CF neona- increase in chromosome detection). Similarly, for
tal screening programs of North America where Hispanic/Mexican CF patients, the ACGM/
trypsinogen/DNA testing is being implemented ACOG panel will detect an estimated 11 muta-
to achieve early diagnosis. The list contains: 1) tions accounting for a 62.8% chromosomal iden-
all mutations present at 0.1% prevalence or tification rate. Our extended panel will encompass
greater, based on analysis of 25,048 U.S. CF chro- 19 mutations in this group and extend the detec-
mosomes from the CFF registry [Cystic Fibrosis tion rate to 72.7% (a 9.9% increase in detection).
Foundation, 1998]; 2) CFTR alleles that have Currently, it is premature to predict the implica-
been commonly identified in many minority tions of this panel on detection in the Native
populations of the United States (African Ameri- American, Arab, and Asian populations that re-
can (12.3% population), Hispanic (12.5% popu- side in the United States, as their population
lation), Native American (0.9% population), specific mutations are poorly defined and char-
and Asian American (3.6% population)) [US acterized at this time.
Census Bureau, 2000b]; and 3) common Euro- The practical advantages of a pan-ethnic
pean founder mutations that are anticipated to mutation panel have recently been described
be common throughout the United States. The in detail for U.S. populations [Heim et al., 2001].
Using two CFTR panels detecting either 70 or screening implemented in these states. This is
86 mutations in an assessment of 5,840 CF chro- particularly true in view of a CF incidence that
mosomes, this study demonstrated better detec- is higher [Kosorok et al., 1996] than originally
tion of CF mutations (81.4%) than that achieved predicted [Kulczycki and Schauf, 1974] in this
with the ACMG/ACOG panel (75.3%). Perhaps population. When examining the remainder of
even more significant is their observation of the country, however, it appears that in most
markedly improved sensitivity in non-Caucasian regions these African American mutations would
populations compared to the 25 mutation panel. be common enough to warrant inclusion. There-
Their rationale for a pan-ethnic mutation fore, we include them in the list of Table 2. His-
panel includes comments that we heartily en- panic populations are traditionally concentrated
dorse: It is impractical to expect comprehen- in the southwestern United States and Califor-
sive assessment of the ethnic background of nia where they can account for greater than 60%
every individual referred for testing . . . in many of the total population in these regions. In addi-
cases, individuals are themselves not fully aware tion, Hispanic individuals account for a signifi-
of their ethnic backgrounds. As the constitution cant proportion (1%9%) of the total population
of the U.S. populations continues to change, a in large number of the remaining states. There-
pan-ethnic approach will be essential. fore, Hispanic-specific mutations should also be
In Figure 3, we present the 1990 U.S. census considered in all screening regions, as listed in
data organized by major ethnic groups to help Table 2.
provide a guide for the application of the entire Figure 4 presents the 2000 census data. Un-
or partial sections of the panel of mutations pre- fortunately, national scale graphical illustrations
sented in Table 2. This census information will at the county level were not yet available at the
help serve as a guide in determining the relative time of publication. When comparing the 1990
influence of ethnic sub-populations on the total and 2000 data sets, we see that there is very little
mutation array in a given region. Native Ameri- change in the relative geographical distribution
can populations tend to be isolated into distinct of most of the ethnic groups across the United
regions throughout the United States. In most States; however, the relative percentage of per-
non-concentrated regions of the United States sons belonging to these various ethnic groups
they make up less than one percent of the total has significantly risen in many regions. As the
population, while in the concentrated regions 2000 census data is further tabulated and inte-
they account for 60% and greater. This pattern grated into databases, more detailed ethnic dis-
allows for easier identification and targeting of tribution maps will be available at http://
specific regions using a customized array of Na- factfinder.census.gov/. This information is essen-
tive American CFTR mutations that can be se- tial in designing a high-yield low-cost screening
lected from Tables 1 and 2. Pacific Islander and panel that is to be applied at the state or county
Asian populations tend to be spread across the level.
states. In most regions they account for less than Not depicted in these graphic illustrations
1% of the total population; however, scattered are strong European influences upon the mu-
pockets of 1%3% and 3%10% do exist. Un- tation pool, specifically, those population that
fortunately, these regions have no distinct pat- have their own distinct mutation spectrum. As
tern, and therefore the mutations specific to an example, we will discuss the Slavic influ-
these populations should be considered in all ence throughout the United States. The ma-
screening regions. Currently, these mutations are
poorly defined and we present only those which
are well known to exist in certain Asian popula- FIGURE 3. Distribution of different ethnic groups across
the United States in 1990. This figure represents US Cen-
tions. Further characterization is needed to con- sus Bureau data from the 1990 US census. Data is plot-
fidently predict useful screening protocols in ted at a countywide resolution with included legends. A:
most Asian populations. Because African Ameri- Distribution of African American persons in the United
States in 1990. B: Distribution of Hispanic persons in
cans account for greater than 50% of the total the United States in 1990. C: Distribution of American
population throughout much of the southeast- Indian, Eskimo, and Aleut persons in the United States
in 1990. D: Distribution of Asian and Pacific Islander
ern United States, it is essential to include these persons in the United States in 1990. [US Census Bu-
mutations into any prenatal or neonatal CF reau, 1990].
FIGURE 4. Distribution of different ethnic groups and total population across the United States in 2000. This figure
represents US Census Bureau data from the 2000 US census. Data is plotted at a statewide resolution with included
legends. A: Distribution of African-American persons. B: Distribution of Hispanic persons. C: Distribution of American
Indian and Alaskan natives. D: Distribution of Asian persons. E: Distribution of Native Hawaiians and other Pacific
Islanders. F: Distribution of total U.S. population in 2000 by state [US Census Bureau, 2000a].
jority of Slavic immigrants came to the United Although most immigrants have assimilated
States between 1880 and 1920, and settled in with the U.S. majority population, strong en-
roughly one-third of the states, focused mainly dogamous central and eastern European popu-
in the Northeast, Midwest, and California lations remain in New York, Toronto, and
(www1.umn.edu/ihrc/sites.htm, www.feefhs.org). Chicago, and are continuously increased by
influx from Poland, Russia, and Romania. Thus, A good example of this hidden bias would be
inclusion of CFTR dele2,3 (21 kb) into the U.S. in Hispanic CF newborns in the United States.
screening panels will substantially improve ge- The 3849+10kbC→T mutation is rare when
netic diagnosis of CF in European Americans compared to the most common CF mutations
of central and eastern European origin. worldwide; however, it represents nearly 2.5%
We have also recommended the inclusion of of the Southwest U.S. Hispanic CF chromo-
other prototypical European mutations, includ- somes. If this mutation is missed in a screening
ing prevalent mutations from Irish, French, Ger- program, the affected CF newborn is at even
man, Scandinavian, Swiss, Welsh, and other greater risk for long-term complications than
strong influences to the American gene pool most CF newborns. This mutation often does
(Table 2). not cause changes in the sweat chloride level,
so the gold standard for clinical diagnosis may
CONCLUSIONS be virtually useless. Furthermore, the mere label
Currently, CF neonatal screening programs of Hispanic or Mexican American tends to di-
are emerging with a variety of methods being minish consideration of CF as a possible diag-
used to facilitate early diagnosis [Hammond et nosis, since CF is typically thought of as a
al., 1991; Wilcken et al., 1995; Gasparini et al., Caucasian or western European disease. Clearly,
1999]. The evolution of tests is described in de- this scenario presents a problem to the diagno-
tail elsewhere [Farrell, 2000] and has featured a sis of CF in certain populations. With the com-
progression from immunoreactive trypsinogen bined efforts of characterizing mutational
testing [Hammond et al., 1991] to a two-tier heterogeneity as a function of geography and
combination of trypsinogen and DNA analysis ethnicity, and the expansion of screening proto-
for either the ∆F508 allele or multiple CFTR cols to account for these population-specific
mutations. With the opportunity for DNA- mutations, many of these missed diagnoses can
based, multi-mutational diagnostics becoming be potentially avoided. Results of the Wiscon-
available, now is the time to initiate planning sin randomized controlled trial have demon-
for uniform and comprehensive newborn screen- strated nutritional benefits [Farrell, 2000] and
ing programs that avoid disparities in detection. other studies suggest better pulmonary outcomes
Many common genetic diseases have been with early diagnosis [Waters et al., 1999; Mérelle
shown to have significant diversity in causative et al., 2001].
mutations when examined on a large population In conclusion, the frequency of most CFTR
or international levels. Cystic fibrosis is perhaps mutations is highly variable and is often a func-
the epitome of mutational diversity since more tion of the ethnic or geographic origin of the
than 1,000 CFTR alleles have been identified parents and grandparents of the affected child.
to date [Cystic Fibrosis Genetic Analysis Con- As new technology is introduced that will allow
sortium (CFGAC), 2000]. With such a great het- for the implementation of large-scale, multi-
erogeneity, those screening programs that rely mutation screening programs, it is essential to
on the detection of a small sub-population of have a complete understanding of what popula-
mutations are almost destined to have an inter- tions are carriers for which mutations. This is
nal bias. Incorporation of such a methodological not only true for cystic fibrosis, but for other
bias into screening programs could significantly common genetic disorders with a large number
reduce the number of newborns identified as CF of causative mutations. Without thorough de-
positive during the genetic tier of screening. lineation of the identity and differential distri-
Therefore, the greatest current challenge in CF bution of mutations at the planning stage for
neonatal screening, when it is employed as a di- newborn screening programs, regions will be
agnostic tool, is to ensure high sensitivity (ide- unprepared to utilize future DNA interrogation
ally 100%) while avoiding exclusion of any technologies as the powerful diagnostic tools that
minority populations. Meeting this challenge will they promise to be, i.e., realize their potential.
require carefully planned trypsinogen/CFTR There is also a great risk of introducing an inter-
multi-mutation analysis [Farrell, 2000] includ- nal bias into any screening program that does
ing the alleles that occur in the population en- not properly select the correct mutational array
compassed by the screening program. associated with the population or populations
that it is servicing. An example of this problem Bonizzato A, Bisceglia L, Marigo C, Nicolis E, Bombieri C,
is provided by the current issue being addressed Castellani C, Borgo G, Zelante L, Mastella G, Cabrini G,
Gasparini P, Pignatti PF. 1995. Analysis of the complete
by the American College of Medical Genetics coding region of the CFTR gene in a cohort of CF patients
(ACMG) and the American College of Obste- from North-East Italy: identification of 90% of the muta-
tricians and Gynecologists (ACOG) regarding tions. Hum Genet 95:397402.
carrier screening [Grody et al., 2001]. The Cabello GMK, Moreira AF, Horovitz D, Correia P, Santa Rosa
ACMG focused its recommendations for the CF A, Llerena J Jr, Greg J, Grody WW, Degrave WM,
carrier screening on non-Jewish Caucasians and Fernandes O, Cabello PH. 1999. Cystic fibrosis: low fre-
Ashkenazi Jews. Several minority populations quency of DF508 mutation in 2 population samples from
Rio de Janeiro, Brazil. Hum Biol 71:189196.
in North America are not well served when only
25 CFTR alleles are included, as Table 2 clearly Cabello GMK, Cabello PH, Llerena J Jr, Fernandes O, Harris
shows. In fact, the panel may not even serve Jew- A. 2001. The 3120+1G→A splicing mutation in CFTR
is common in Brazilian cystic fibrosis patients. Hum Biol-
ish populations adequately based on a recent ogy 73:403409.
study aimed at evaluation of screening policies
Carles S, Desgeorges M, Goldman A, Thiart R, Guittard C,
[Orgad et al., 2001]. In addition, another study Kitazos CA, de Ravel TJL, Westwood ART, Claustres M,
of a typical CF Center population of ethnically Ramsay M. 1996. First report of CFTR mutations in black
diverse patients residing in California under- cystic fibrosis patients of southern African origin. J Med
scores the need for a pan-ethnic mutation Genet 33:802804.
panel [Heim et al., 2001] in a neonatal screen- Cartault F, Steffann J, Vidaud D, Bousquet S, Lesure F, Renouil
ing program [Wine et al., 2001]. Consequently, M, McDonell N, Feingold J, Beldjord C, Bienvenu T. 1996.
ultimate goals for any CF screening program Detection of more than 91% cystic fibrosis mutations in a
sample of the population from Reunion Island and identi-
must first be identification of the population- fication of two novel mutations (A309G, S1255L) and one
specific CF mutational arrays, secondly the ex- novel polymorphism (L49L). Clin Genet 54:437439.
pansion of counseling and support networks, and
Casals T, Nunes V, Palacio A, Giménez J, Gaona A, Ibáñez N,
thirdly, the implementation of a population-tai- Morral N, Estivill X. 1993. Cystic fibrosis in Spain: high
lored screening protocol. With the information frequency of mutation G542X in the Mediterranean coastal
provided in this article and using recommended area. Hum Genet 91:6670.
procedures [Farrell, 2000], newborn screening Casals T, Ramos MD, Giménez J, Larriba S, Nunes V, Estivill
for CF can soon progress from the current IRT/ X. 1997. High heterogeneity for cystic fibrosis in Spanish
DNA (∆F508) methods to a superior IRT/DNA families: 75 mutations account for 90% of chromosomes.
(CFTR) strategy. Hum Genet 101:365370.
Cashman SM, Patino A, Martinez A, Garcia-Delgado M,
REFERENCES Miedzbrodzka Z, Schwarz M, Shrimpton A, Ferec C,
Ranguenes O, Macek M. 1995. Identical intragenic
Allen DB, Farrell PM. 1996. Newborn screening. Adv Pediatr
microsatellite haplotype found in cystic fibrosis chromo-
43:231270. somes bearing mutation G551D in Irish, English, Scottish,
Angelicheva D, Calafell F, Savov A, Jordanova A, Kufardjieva Breton and Czech patients. Hum Hered 45:612.
A, Galeva I, Nedkova V, Ianova T, Yankova P, Kon- Castaldo G, Rippa E, Sebastio G, Raia V, Ercolini P, de Ritis G,
stantinova D, Genev E, Kalaydjieva L. 1997. Cystic fibro- Salvatore D, Slavatore F. 1996. Molecular epidemiology of
sis mutations and associated haplotypes in Bulgariaa cystic fibrosis mutations and haplotype in southern Italy
comparative population genetic study. Hum Genet 99: evaluated with an improved semiautomatic robotic pro-
513520. cedure. J Med Genet 33:475479.
Banjar H, Kambouris M, Meyer BF, Al-Mehaidib A, Mogarri Castaldo G, Fuccio A, Cazeneuve C, Picci L, Salvatore D,
I. 1999. Geographic distribution of cystic fibrosis trans- Raia V, Scarpa M, Goossens M, Salvatore F. 1999. Detec-
membrane regulator gene mutations in Saudi Arabia. Ann tion of five rare cystic fibrosis mutations peculiar to South-
Trop Paediatr 19:6973. ern Italy: implications in screening for the disease and
Bayleran JK, Yan H, Hopper CA, Simpson EM. 1996. Fre- phenotype characterization for patients with homozygote
quencies of cystic fibrosis mutations in the Maine popula- mutations. Clin Chem 45:957962.
tion: high proportion of unknown alleles in individuals of Cheadle JP, Goodchild MC, Meredith Al. 1993. Direct se-
French-Canadian ancestry. Hum Genet 98:207209. quencing of the complete CFTR gene: the molecular char-
Bernardino ALF, Ferri A, Passos-Bueno MR, Kim CEA, Nakaie acterization of 99.5% of CF chromosomes in Wales. Hum
CMA, Gomes CET, Damaceno N, Zatz M. 2000. Molecu- Molec Genet 2:15511556.
lar analysis in Brazilian cystic fibrosis patients reveals five Chertkoff L, Visich A, Bienvenu T, Grenoville M, Segal E,
novel mutations. Genet Test 4:6974. Carniglia L, Kaplan JC, Barreiro C. 1997. Spectrum of
CFTR mutations in Argentine cystic fibrosis patients. Clin gene of German cystic fibrosis patients. Hum Genet
Genet 51:4347. 90:279284.
Chevalier-Porst F, Bonardot AM, Gilly R, Chazalette JP, Dörk T, Mekus F, Schmidt K, Boßhammer J, Fislage R,
Mathieu M, Bozon D. 1994. Mutation analysis in 600 Heuer T, Dziadek V, Neuman T, Kälin N, Wulbrand U,
French cystic fibrosis patients. J Med Genet 31:541544. Wulf B, von der Hardt H, Maaß G, Tümmler B. 1994.
Detection of more than 50 different CFTR mutations
Chillón M, Casals T, Giménez J, Ramos MD, Palacio A, Morral
N, Estivill X, Nunes V. 1994. Analysis of the CFTR gene in a large group of German cystic fibrosis patients. Hum
confirms the high genetic heterogeneity of the Spanish Genet 94:542553.
population: 43 mutation account for only 78% of CF chro- Dörk T, El-Harith EA, Stuhrmann M, Macek M Jr, Egan M,
mosomes. Hum Genet 93:447451. Cutting GR, Tzetis M, Kanavakas E, Carles S, Claustres
Claustres M, Lassel M, Desgeorges M, Giansily M, Culard JF, M, Padoa C, Ramsay M, Schmidke J. 1998. Evidence for a
Razakatsara G, Demaille J. 1993. Analysis of the 27 exons common ethnic origin of cystic fibrosis mutation
and flanking regions of the cystic fibrosis gene: 40 differ- 3120+1G→A in diverse populations. Am J Hum Genet
ent mutations account for 91.2% of the mutant alleles in 63:656662.
Southern France. Hum Molec Genet 2:12091213. Dörk T, Macek M Jr, Mekus F, Tümmler B, Tzountzouris J,
Claustres M, Guittard C, Bozon D, Chevalier F, Verlingue C, Casals T, Krebsová A, Koudová M, Sakmaryová I, Macek
M Sr, Várová V, Zemková D, Ginter E, Petrova N,
Ferec C, Girodon E, Cazenueve C, Bienvenu T, Lalau G,
Ivaschenko T, Baranov V, Witt M, Pogorzelski A, Bal J,
Dumur V, Feldmann D, Bieth E, Blayau M, Clavel C, Zékanowsky C, Wagner K, Stuhrmann M, Bauer I,
Creveaux I, Malinge M, Monnier N, Malzac P, Mittre H, Seydewitz HH, Neumann T, Jakubiczka S, Kraus C,
Chomel J, Bonnefont J, Iron A, Chery M, Des Georges M. Thamm B, Nechiporenko M, Livshits L, Mosse N,
2000. Spectrum of CFTR mutations in cystic fibrosis and Tsukerman G, Kadási L, Ravnik-Glavac M, Glavac D,
in congenital absence of the vas deferens in France. Hum Komel R, Vouk K, Kucinskas V, Krumina A, Teder M,
Mutat 16:143156. Kocheva S, Efremov GD, Onay T, Kirdar B, Malone G,
Collee JM, de Vries HG, Scheffer H, Halley DJJ, ten Kate LP. Schwarz M, Zhou Z, Friedman KJ, Carles S, Claustres M,
1998. Relative frequencies of cystic fibrosis mutations in Bozon D, Verlingue C, Férec C, Tzets M, Kanavakis E,
The Netherlands as an illustration of significant regional Cuppens H, Bombieri C, Pignatti PF, Sangiuolo F,
variation in a small country. Hum Genet 102:587590. Jordanova A, Kusic J, Radojkovic D, Sertic J, Richter D,
Rukavina AS, Bjorck E, Strandvik B, Cardoso H, Mont-
Cuppens H, Marynen P, De Boeck C, Cassiman JJ. 1993. De- gomery M, Nakielna B, Hughes D, Estivill X, Aznarez I,
tection of 98.5% of the mutations in 200 Belgian cystic Tullis E, Tsui LC, Zielenski J. 2000. Characterization of a
fibrosis alleles by reverse dot-blot and sequencing of the novel 21kb deletion, CFTRdele2,3(21 kb), in the CFTR
complete coding region and exon/intron junctions of the gene a cystic fibrosis mutation of Slavic origin common in
CFTR gene. Genomics 18:693697. Central and East Europe. Hum Genet 106:259268.
Cystic Fibrosis Foundation. 1998. Patient registry, 1997: an- El-Harith EA, Dörk T, Stuhrmann M, Abu-Srair H, Al-Shahri
nual data report. Bethesda MD: Cystic Fibrosis Foun- A, Keller KM, Lentze MJ, Schmidtke J. 1997. Novel and
dation. p. 3. characteristic CFTR mutations in Saudi Arab children with
Cystic Fibrosis Genetic Analysis Consortium (CFGAC). 1994. severe cystic fibrosis. J Med Genet 34:996999.
Population variation of common cystic fibrosis mutations.
Estivill X, Bancells C, Ramos C, Biomed CF Mutation Analy-
Hum Mutat 4:167177.
sis Consortium. 1997. Geographic distribution and regional
Cystic Fibrosis Genetic Analysis Consortium (CFGAC). 2000. origin of 272 cystic fibrosis mutations in European popula-
www.genet.sickkids.on.ca/cftr/. tions. Hum Mutat 10:135154.
Dabovic BB, Radojkovic D, Minic P, Savic J, Savic A. 1992. European Working Group on CF Genetics. 1990. Gradient of
Frequency of the ∆F508 deletion and G551D, R553X and distribution in Europe of the major CF mutation and its
G542X mutations in Yugoslav CF patients. Hum Genet associated haplotypes. Hum Genet 85:436445.
88:699700.
Farrell PM. 2000. Improving the health of patients with cystic
De Braekeleer M, Mari C, Verlingue C, Allard C, Leblanc fibrosis through newborn screening. Adv Ped 49:79115.
JP, Simard F, Aubin G, Férec C. 1998. Complete iden-
tification of cystic fibrosis transmembrane conductance Férec C, Audrezet MP, Mercier B, Guillermit H, Moullier P,
regulator mutations in the CF population of Saguenay Quere I, Verlingue C. 1992. Detection of over 98% cystic
Lac-Saint-Jean (Quebec, Canada). Clin Genet 53: fibrosis mutations in a Celtic population. Nat Genet
4446. 1:188191.
Desgeorges M, Mégarbané A, Guittard C, Carles S, Loiselet J, Friedman KJ, Leigh MW, Czarnecki P, Feldman GL. 1998.
Demaille J, Claustres M. 1997. Cystic fibrosis in Lebanon: Cystic fibrosis transmembrane-conductance regulator
distribution of CFTR mutations among Arab communi- mutations among African Americans. Am J Hum Genet
ties. Hum Genet 100:279283. 62:195196.
Dörk T, Kälin N, Stuhrmann M, Schmidtke J, Tümmler B. Frossard PM, Girodon E, Dawson KP, Ghanem N, Plassa F,
1992. A termination mutation (2143delT) in the CFTR Lestringant GG, Gossens M. 1998. Identification of cystic
fibrosis mutations in the United Arab Emirates. Hum James S. 1993. The world of the Celts. London: Thames and
Mutat 1:412413. Hudson, Ltd.
Frossard PM, Lestringant G, Girodon E, Goosens M, Dawson Kambouris M, Banjar H, Moggari I, Al Hamed MH, El Naser
KP. 1999. Determination of the prevalence of cystic fibro- H, Meyer BF. 1997. Identification of novel mutations in
sis in the United Arab Emirates by genetic carrier screen- the CFTR gene causing cystic fibrosis (CF) in Arab popu-
ing. Clin Genet 55:496497. lations. Institut Paster del Tunis 74:9394.
Gan KH, Veeze HJ, Van Den Ouweland AMW, Halley DJJ, Kanavakis E, Tzetis M, Antoniadi T, Traeger-Syodinos J,
Scheffer H, Van Der Hout A, Overbeek SE, De Jongste Doudounakis S, Adam G, Matsaniotis N, Kattamis C.
JC, Bakker W, Heijerman HGM. 1995. A cystic fibrosis 1995. Mutation analysis of ten exons of the CFTR gene
mutation associated with mild lung disease. NEJM in Greek cystic fibrosis patients: characterization of 74.5%
333:9599. of CF alleles including one novel mutation. Hum Genet
Gasparini P, Arbustini E, Restagno G, Zelante L, Stanziale P, 96:364366.
Gatta L, Sbaiz L, Sedita AM, Banchieri N, Sapone L, Fiorucci Kere J, Estivill X, Chillón M, Morral N, Nunes V, Norio R,
GC, Brison E, Shulse E, Rappaport E, Fortina P. 1999. Analy- Savilahti E, de la Chillón M. 1994. Cystic fibrosis in a low-
sis of 31 CFTR mutations by polymerase chain reaction/oli- incidence population: two major mutations in Finland.
gonucleotide ligation assay in a pilot screening of 4476 Hum Genet 93:162166.
newborns for cystic fibrosis. L Med Screen 3:6769.
Kerem B, Rommens JM, Buchanan JA, Markiewicz D, Cox
Gilfillan A, Warner JP, Kirk JM, Marshall T, Greening A, Ho TK, Chakravarti A, Buchwald M, Tsui LC. 1989. Identifi-
LP, Hargreave T, Stack B, McIntyre D, Davidson R, Dean cation of the cystic fibrosis gene: genetic analysis. Science
JCS, Middleton W, Brock DJH. 1998. P67L: a cystic fibro- 245:10731080.
sis allele with mild effects found at high frequency in the
Scottish population. J Med Genet 35:122125. Kerem E, Kalman Y, Yahav Y, Shoshani T, Abeliovich C,
Szeinberg A, Rivlin J, Blau H, Tal A, Ben-Tur L, Spinger
Goldman A, Labrum R, Claustres M, Desgeorges M, Guittard C, Augarten A, Godfrey S, Lerer I, Branski D, Friedman
C, Wallace A, Ramsay M. 2001. The molecular basis of M, Kerem B. 1995. Highly variable incidence of cystic fi-
cystic fibrosis in South Africa. Clin Genet 59:3741. brosis and different mutation distribution among different
Grebe TA, Seltzer WK, DeMarchi J, Silva DK, Doane WW, Jewish ethnic groups in Israel. Hum Genet 96:193197.
Gozal D, Richter SF, Bowman CM, Norman RA, Rhodes
Klaassen T, Teder M, Viikmaa M, Metspala A. 1998. Neona-
SN, Hernried LS, Murphy S, Harwood IR, Accurso FJ, Jain
tal screening for the cystic fibrosis main mutation ∆F508
KD. 1994. Genetic analysis of Hispanic individuals with
in Estonia. J Med Screen 5:1619.
cystic fibrosis. Am J Hum Genet 54:443446.
Gregg RG, Simantel A, Farrell PM, Koscik RE, Kosorok MR, Kosorok MR, Wei W, Farrel PM. 1996. The incidence of cys-
Laxova A, Laessig RH, Hoffman GL, Hassemer DJ, tic fibrosis. Stat Med 15:449462.
Mischler EH, Splaingard M. 1997. Newborn screening for Kristidis P, Bozon D, Corey M, Markiewicz D, Rommens J,
cystic fibrosis in Wisconsin: comparison of biochemical and Tsui LC, Durie P. 1992. Genetic determination of exocrine
molecular methods. Pediatrics 99:819824. pancreatic function in cystic fibrosis. Am J Hum Genet
Grody WW, Cutting GR, Klinger KW, Richards CS, Watson 50:11781184.
MS, Desnick RJ. 2001. Laboratory standards and guide- Kulczycki LL, Schauf V. 1974. Cystic fibrosis in blacks in Wash-
lines for population based cystic fibrosis screening. Genet ington, DC. AmJ Dis Child 127:6467.
Med 3:149154.
Liang MH, Wertz KK, Bowman CM, Hsu E, Shapiro B, Wong
Guilloud-Bataille M, De Crozes D, Rault G, Degioanni A, LC. 1998. 3849+10kbC→T splicing mutation in Hispanic
Feingold J, The Clinical Centers of the CF French Regis- CF patients. Mol Genet Metab 64:213216.
try. 2000. Cystic fibrosis mutations: report from the French
registry. Hum Hered 50:142145. Loirat F, Hazout S, Lucotte G. 1997. G542X as a probable
Phoenician cystic fibrosis mutation. Hum Biol 69:419425.
Hammond KB, Abman SH, Sokol RJ, Accurso FK. 1991. Ef-
ficacy of statewide neonatal screening for cystic fibrosis by Loumi O, Baghriche M, Delpech M, Kaplan JC, Bienvenu T.
assay of trypsinogen concentrations. N Engl J Med 1999. Analysis of the complete coding region of the CFTR
325:769774. gene in ten Algerian cystic fibrosis families. Hum Hered
49:8184.
Heim R, Sugarman MS, Allitto BA. 2001. Improved detec-
tion of cystic fibrosis mutations in the heterogeneous U.S. Lucotte G, Hazout S, De Braekeleer M. 1995. Complete map
population using an expanded, pan-ethnic mutation panel. of cystic fibrosis mutation DF508 frequencies in Western
Genet Med 3:168176. Europe and correlation between mutation frequencies and
incidence of disease. Hum Biol 67:797803.
Hergersberg M, Balakrishnan J, Bettecken T, Bettecken T,
Chevalier-Porst F, Brägger C, Burger R, Liechti-Gallati S, Macek M Jr, Macek M, Serre JL, Vavrova V, Burger J, Reis A,
Moris M, Schorderet D, Thonney F, Moser H, Malik N. Schmidke J, Sperling K, Cutting G. 1991. Population study
1997. A new mutation, 3905insT, accounts for 4.8% of of CFTR gene mutations in Bohemia and Moravia: hy-
1173 CF chromosomes in Switzerland and causes a severe pothesis on the historical spread of G551D and deltaF508
phenotype. Hum Genet 100:220223. in Europe. Am J Hum Genet 49:A474, A2703.
Macek M Jr, Mackova A, Hamosh A, Hilman BC, Selden sociated haplotypes in Turkish cystic fibrosis patients. Hum
RF, Lucotte G, Friedman KJ, Knowles MR, Rosenstein Biology 73:191203.
BJ, Cutting GR. 1997. Identification of common cystic
Orgad S, Neumann S, Loewenthal R, Netanelov-Shapira I,
fibrosis mutations in African-Americans with cystic fi-
Gazit E. 2001. Prevalence of cystic fibrosis mutations in
brosis increases the detection rate to 75%. Am J Hum
Israeli jews. Genet Testing 5:4752.
Genet 60:11221127.
Orozco L, Salcedo M, Lezana JL, Chvez M, Valdez H, Moreno
Macek M Jr et al. 2002. Eur J Hum Genet (in press). M, Carnevale A. 1993. Frequency of ∆F508 in a Mexican
Malone G, Haworth A, Schwartz MJ, Cuppens H, Super sample of cystic fibrosis patients. J Med Genet 30:501502.
M. 1998. Detection of five novel mutations of the cys- Orozco L, Velázquez R, Zielenski J, Tsui LC, Chávez M, Lezana
tic fibrosis transmembrane regulator (CFTR) gene in Pa- JL, Saldaña Y, Hernández E, Carnevale A. 2000. Spec-
kistani patients with cystic fibrosis: Y569D, Q98X, trum of CFTR mutation in Mexican cystic fibrosis patients:
296+12(T>C), 1161delC and 621+2(T>C). Hum identification of five novel mutations (W1098C, 846delT,
Mutat 11:152157. P750L, 4160insGGGG and 297-1G→A). Hum Genet
Mercier B, Mac Clure M, Grebe T, Doane WW, Audrézet MP, 106:360365.
Quéré I, Kaplan GC, Férec C. 1992. Screening for CFTR Parizotto EA, Bertuzzo CS. 1997. Molecular characterization
gene mutations in a Southwest American Indian popula- of cystic fibrosis patients in the state of São Paulo (Brazil).
tion. Am J Hum Genet 51(suppl):A340; 1341. J Med Genet 34:877.
Mercier B, Lissens W, Audrezet MP, Bonuelle M, Liebaers I, Paz-y-Mino C, Pérez JC, Burgos R, Dávalos Ma V, Leone PE.
Ferec C. 1993. Detection of more than 94% cystic fibrosis 1999. The ∆F508 mutation in Ecuador, South America.
mutations in a sample of Belgian population and identifi- Hum Mutat 14:348350.
cation of four novel mutations. Hum Mutat 2:1620.
Petreska L, Koceva S, Plaseska D, Chernick M, Gordova-
Mercier B, Raguénès O, Estivill X, Morral N, Kaplan GC, Muratovska A, Fustic S, Nestorov R, Nestorov R, Efremov
McClure M, Grebe TA, Kessler D, Pignatti PF, Marigo C, GD. 1998. Molecular basis of cystic fibrosis in the Repub-
Bombieri C, Audrezet MP, Verlingue C, Férec C. 1994. lic of Macedonia. Clin Genet 54:203209.
Complete detection of mutations in cystic fibrosis patients
Popa I, Pop L, Popa Z, Schwarz MJ, Hambleton G, Malone
of Native American origin. Hum Genet 94:629632.
GM, Haworth A, Super M. 1997. Cystic fibrosis mutations
Mérelle ME, Schouten JP, Gerritsen J, Dankert-Roelse JE. in Romania. Eur J Pediatr 156:212213.
2001. Influence of neonatal screening and centralized treat- Raskin S, Phillips JA, Kaplan G, McClure M, Vnencak-Jones
ment on long-term clinical outcome and survival of CF C, Rozov T, Cardieri JM, Marostica P, Abreu F, Giugliani
patients. Eur Respir J 18:306315. R, Reis F, Rosario NA, Ludwig N, Periera L, Faucz F,
Messaoud T, Verlingue C, Denamur E, Pascaud O, Quéré I, Gabardo J, Culpi L. 1999. Geographic heterogeneity of 4
Fattoum S, Elion J, Férec C. 1996. Distribution of CFTR common worldwide cystic fibrosis non-DF508 mutations
mutations in cystic fibrosis patients of Tunisian origin: in Brazil. Hum Biol 71:111121.
identification of two novel mutations. Euro J Hum Genet Restrepo CM, Pineda L, Rojas-Martínez A, Gutiérrez CA,
4:2024. Morales A, Gómez Y, Villalobos MC, Borjas L, Delgado
Morales MM, Capella MAM, Lopes AG. 1999. Structure and W, Myers A, Barrera-Saldaña HA. 2000. CFTR muta-
function of the cystic fibrosis transmembrane conductance tions in three Latin American countries. Am J Med Genet
regulator. Braz J Med Biol Res 32:10211028. 91:277279.
Morral N, Bertranpetit J, Estivell X, Nunes V, Casals T, Riordan JR, Rommens JM, Kerem B, Alon N, Rozmahel R,
Giménez J, Reis A, Varon-Mateeva R, Macek M Jr, Grzelczak Z, Zielinski J, Lok S, Plavsic N, Chou JL, Drumm
Kalaydjieva L, Angelcheva D, Dancheva R, Romeo G, ML, Iannuzzi MC, Collins FS, Tsui LC. 1989. Identifica-
Russo MP, Garnerone S, Restango G, Ferrari M, Magnani tion of the cystic fibrosis gene: cloning and characteriza-
C, Claustres M, Desgeorges M, Schwarz M, Dallapiccola tion of complementary DNA. Science 245:10661072.
B, Novelli G, Férec C, de Arce M, Nemeti M, Kere J, Rios J, Orellana O, Aspillaga M, Avendano I, Largo I, Riveros
Anvret M, Dahl N, Kadasi L. 1994. The origin of the N. 1994. CFTR mutations in Chilean cystic fibrosis pa-
major cystic fibrosis mutation (∆F508) in European tients. Hum Genet 94:291294.
populations. Nat Genet 7:169175.
Romeo G, Devoto M, Galietta LJV. 1989. Why is the cystic
Onay T, Topaloglu O, Zielenski J, Gokgoz N, Kayserili H, fibrosis gene so frequent? Hum Genet 84:15.
Camcioglu Y, Cokugras H, Akcakaya N, Apak M, Tsui
LC, Kirdar B. 1998. Analysis of the CFTR gene in Turk- Rommens JM, Iannuzzi MC, Kerem B, Drumm ML, Melmer
ish cystic fibrosis patients: identification of three novel G, Dean M, Rozmahel R, Cole JL, Kennedy D, Hidaka N,
mutations (3172delAC, P1013L and M1028I). Hum Zsiga M, Buchwald M, Riordan JR, Tsui LC, Collins FS.
Genet 102:224230. 1989. Identification of the cystic fibrosis gene: chromo-
some walking and jumping. Science 245:10591065.
Onay T, Zielenski J, Topaloglu O, Gokgoz N, Kayseril H,
Yukselapak M, Camcioglu Y, Cokugras H, Akcakaya N, Rozen R, De Braekeleer M, Daigneault J, Ferreira-Rajabi L,
Tsu LC, Kirdar B. 2001. Cystic fibrosis mutations as as- Gerdes M, Lamoureux L, Aubin G, Simard F, Fujiwara TM,
Morgan K. 1992. Cystic fibrosis mutations in French Ca- cystic fibrosis patient with the 1998+1G→T splicing mu-
nadians: three CFTR mutations are relatively frequent in tation in the CFTR gene: a review of East Asian cases.
a Quebec population with an elevated incidence of cystic Hum Mutat 12:361.
fibrosis. Am J Med Genet 42:360364.
Tümmler B, Storrs T, Dziadek V, Dörk T, Meitinger T, Golla
Schaedel C, Hjelte L, de Monestrol I, Johannesson M, Kollberg A, Bertele-Harms RM, Harms HK, Schröder E, Claaß A,
H, Kornfält R, Holmberg L. 1999. Three common CFTR Rutjes J, Schneppenheim R, Bauer I, Breuel K, Stuhrmann
mutations should be included in a neonatal screening M, Schmidtke J, Linder M, Eigel A, Horst J, Kaiser R,
programme for cystic fibrosis in Sweden. Clin Genet Lentze MJ, Schmidt K, von der Hardt H, Estivill X. 1996.
56:318322. Geographic distribution and origin of CFTR mutations in
Germany. Hum Genet 97:727731.
Schwartz M, Anvret M, Claustres M, Geir Eiken H, Eiklid K,
Schaedel C, Stolpe L, Tranebjærg L. 1994. 394delTT: a Tzetis M, Kanavakis E, Antoniadi T, Doudounakis S, Adam
Nordic cystic fibrosis mutation. Hum Genet 93:157161. G, Kattamis C. 1997. Characterization of more than 85%
of cystic fibrosis alleles in the Greek population, including
Schwartz M, Sørensen N, Brandt NJ, Høgdall E, Holm T. five novel mutations. Hum Genet 99:121125.
1995a. High incidence of cystic fibrosis on the Faroe Is-
lands: a molecular and genealogical study. Hum Genet US Census Bureau. 1990. Race and Hispanic origin popula-
95:703706. tion density of the United States: 1990. www.census.gov/
geo/www/mapGallery/RHOriginPD-1990.html.
Schwartz MJ, Malone GM, Haworth A, Cheadle JP, Meredith
L, Gardner A, Sawyer IH, Connarty M, Dennis N, Seller US Census Bureau. 2000a. American fact finder: thematic
A, Harris A, Taylor R, Dear S, Middleton-Price H, maps section. http://factfinder.census.gov/.
McMahon C, Mayall E, McMahon R, Barton DE, Giles US Census Bureau. 2000b. Census of population and hous-
M, Lindley V, Plaha DS, Price S, Sharif A, Cross GS, Dalton ing: 2000. www.census.gov/prod/cen2000/index.html.
A, Taylor G, Wallace A, Tassabehji M, Whittaker JL, But-
ler R, Curtis A, Pinkett R, Gilfillan AJ, Brock DJH, Higgins Villalobos-Torres C, Rojas-Martínez A, Villareal-Castellanos
GS, Lanyon G, Miedzybrodzka Z, Davidson M, Graham E, Cantú JM, Sánchez-Anxaldo FJ, Saiki RK, Barrera-
Saldaña HA. 1997. Analysis of 16 cystic fibrosis
CA, Hill AJM. 1995b. Cystic fibrosis mutation analysis:
mutations in Mexican patients. Am J Med Genet 69:
report from 22 U.K. regional genetics laboratories. Hum
Mutat 6:326333. 380382.
Waters DL, Wilcken B, Irwig L, Van Asperen P, Mellis C,
Scotet V, De Braekeleer M, Roussey M, Rault G, Parent P,
Simpson JM, Brown J, Gaskin KJ. 1999. Clinical outcomes
Dagorne M, Journel H, Lemoigne A, Codet J, Catheline
of newborn screening for cystic fibrosis. Arch Dis Child
M, David V, Chaventré A, Duguépéroux I, Verlingue C,
Fetal Neonatal Ed 80:F1F7.
Quéré I, Mercier B, Audrézet M, Férec C. 2000. Neonatal
screening for cystic fibrosis in Brittany, France: assessment Welsh MJ, Tsui L, Boat TF, Beaudet AL. 1995. The metabolic
of 10 years experience and impact on prenatal diagnosis. and molecular bases of inherited disease. 17th ed. Cystic
Lancet 356:789794. fibrosis. New York: McGraw-Hill. p 37993876.
Serre JL, Simon-Bouy B, Mornet E, Jaume-Roig B, Basso- Wilcken D, Wiley V, Sherry G, Bayliss U. 1995. Neonatal screen-
poulou A, Schwartz M, Taillandier A, Boué A. 1990. Stud- ing for cystic fibrosis: a comparison of two strategies for case
ies of RFLP closely linked to the cystic fibrosis locus through detection in 1.2 million babies. J Pediatr 127:965970.
Europe lead to new considerations in population genetics. Wine JJ, Kuo E, Hurlock G, Moss RB. 2001. Comprehensive
Hum Genet 84:449454. mutation screening in a cystic fibrosis center. Pediatrics
Shoshani T, Augarten A, Gazit E, Bashan N, Yahav Y, Rivlin 107:280286.
Y, Tal A, Seret H, Yaar L, Kerem E, Kerem B. 1992. Asso- Yee K, Robinson C, Hurlock G, Moss R, Wine JJ. 2000. Novel
ciation of a nonsense mutation (W1282X), the most com- cystic fibrosis mutation L1093P: functional analysis and
mon mutation in the Ashkenazi Jewish cystic fibrosis possible Native American origin. Hum Mutat 15:208.
patients in Israel, with presentation of severe disease. Am
J Hum Genet 50:222228. Yilmaz E, Erdem H, Özgüç M, Coskun T, Özçelik U, Göçmem
S, Özalp I. 1995. Study of 12 mutations in Turkish cystic
Shrimpton AE, McIntosh I, Brock JH. 1991. The incidence fibrosis patients. Hum Hered 45:175177.
of different cystic fibrosis mutations in the Scottish popu-
lation: effects on prenatal diagnosis and genetic counsel- Zielenski J, Fujiwara TM, Markiewicz D, Paradis AJ, Anacleto
ing. J Med Genet 28:317321. AI, Richards B, Schwartz RH, Klinger KW, Tsui LC, Mor-
gan K. 1993. Identification of the M1101K mutation in
Stuhrmann M, Dörk T, Frühwirth M, Golla A, Skawran B, the cystic fibrosis transmembrane conductance regulator
Antonon W, Ebhardt M, Loos A, Ellemunter H, Schmidtke (CFTR) gene and complete detection of cystic fibrosis
J. 1997. Detection of 100% of the CFTR mutations in 63 mutations in the Hutterite population. Am J Hum Genet
families from Tyrol. Clin Genet 52:240246. 52:609615.
Suwanjutha S, Huang NH, Wattanasirichaigoon D, Sara T, Zielenski J, Tsui LC. 1995. Cystic fibrosis: genotypic and phe-
Harris A, Macek Jr M. 1998. Case report of a Tai male notypic variations. Ann Rev Genet 29:777807.

Bobadilla 2002

Uploaded by

Copyright:

Available Formats

Bobadilla 2002

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Bobadilla 2002

Uploaded by

Copyright:

Available Formats

HUMAN MUTATION 19:575606 (2002)

Cystic Fibrosis: A Worldwide Analysis of CFTR

Communicated by Lap-Chee Tsui

© 2002 Wiley-Liss, Inc.

INTRODUCTION 1994]. The large number of more common mu-

of mutations were evaluated via linear regres- STATISTICAL IMPLICATIONS

TABLE 3. Filtered Data for CF Incidence Across the Globe*

You might also like

Bobadilla 2002

Uploaded by

Copyright:

Available Formats

Bobadilla 2002

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Bobadilla 2002

Uploaded by

Copyright:

Available Formats

HUMAN MUTATION 19:575606 (2002)

Cystic Fibrosis: A Worldwide Analysis of CFTR

Communicated by Lap-Chee Tsui

© 2002 Wiley-Liss, Inc.

INTRODUCTION 1994]. The large number of more common mu-

of mutations were evaluated via linear regres- STATISTICAL IMPLICATIONS

TABLE 3. Filtered Data for CF Incidence Across the Globe*

You might also like

HUMAN MUTATION 19:575606 (2002)