RESEARCH ARTICLE Open Access
Effects of SLC10A2 variant rs9514089 on gallstone
risk and serum cholesterol levels- meta-analysis
of three independent cohorts
Anke Tönjes1*, Henning Wittenburg2, Jan Halbritter1, Olga Renner3, Simone Harsch3, Eduard F Stange4,
Frank Lammert5, Michael Stumvoll1 and Peter Kovacs6
Abstract
Background: Recently, a single nucleotide polymorphism (SNP) rs9514089 in SLC10A2 (apical sodium-dependent
bile acid transporter gene) has been identified as a susceptibility variant for cholelithiasis in humans.
Methods: Here we assessed the effects of rs9514089 on gallstone risk and related phenotypes of the metabolic
syndrome in the self-contained population of Sorbs (183 cases with gallstones/826 controls). Furthermore, we
performed a meta-analysis for effects of rs9514089 on susceptibility for cholelithiasis in three independent cohorts
(Stuttgart: 56 cases/71 controls, Aachen: 184 cases/184 controls and Sorbs).
Results: There was no significant association of rs9514089 with gallstone risk, serum lipid parameters and BMI in
the Sorbs and in the meta-analysis of all three cohorts (p > 0.05). There was an effect trend in the subgroup of
lean subjects but based on different effect directions in the three cohorts there was no significant association in
the meta-analysis.
Conclusions: We were not able to replicate the effect of rs9514089 on gallstone risk in the Sorbs. Further analyses
in larger cohorts are required to finally assess the role of genetic variants in SLC10A2 in human gallstone
development and lipid metabolism.
Background
The pathogenesis of gallstone disease is complex and a
variety of environmental predisposing factors such as
obesity and rapid weight loss, nutrition, certain medica-
tions and number of pregnancies have been identified
[1-6]. However, human and murine data suggest a
strong genetic component for the risk of gallstone for-
mation [7-18]. Very recently, Renner et al. have identi-
fied SLC10A2 (apical sodium-dependent bile acid
transporter; protein name ASBT) as a novel susceptibil-
ity gene for cholelithiasis in humans [19]. SLC10A2
encodes the cholangiocyte bile salt transporter protein
whose expression is reduced by a lithogenic diet in mice
[20] and mediates intestinal bile acid absorption [21,22].
ASBT is regulated by changes in gene expression in
response to biliary bile salt concentration and inflamma-
tory cytokines and is thought to enable cholangiocytes
to sense biliary bile salts in order to activate intracellular
signaling pathways [23] and to promote cholehepatic
shunting of bile salts. Furthermore, ASBT expression in
the cholangiocyte apical membrane is regulated by
secretin [24]. It has been shown that in gallstone
patients, ileal ASBT expression is diminished, and that
this is associated with low cytosolic ileal lipid binding
protein (ILBP) and basolateral organic solute bile acid
exporter (OSTa-OSTb) expression indicating impaired
enterohepatic circulation of bile salts at least in a subset
of patients with cholelithiasis [25,26]. The regulation of
ASBT expression also appears weight specific and so far,
a diminished ASBT expression has only been shown in
non-obese gallstone patients [25,26].
The recent study by Renner et al. suggested that a sin-
gle nucleotide polymorphism (SNP) rs9514089 mapping
within the SLC10A2 locus is a genetic determinant of
* Correspondence: anke.toenjes@medizin.uni-leipzig.de
1Department of Medicine, Division of Endocrinology and Nephrology,
University of Leipzig, Leipzig, Germany
Full list of author information is available at the end of the article
Tönjes et al. BMC Medical Genetics 2011, 12:149
http://www.biomedcentral.com/1471-2350/12/149
© 2011 Tönjes et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons
Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in
any medium, provided the original work is properly cited.
gallstone disease, expressing gender and weight specifi-
city with higher risk observed in men and in normal-
weight subjects [19]. Since the sample sizes of the two
cohorts from Germany included in the recent study
were rather small (N = 127 from Stuttgart and N = 368
from Aachen), we assessed the effects of rs9514089 on
gallstone risk and related phenotypes of the metabolic
syndrome in the self-contained population of Sorbs and
performed a meta-analysis for effects of rs9514089 on
susceptibility for cholelithiasis in the three independent
cohorts (Stuttgart, Aachen and Sorbs).
Methods
Subjects
All subjects are part of a sample from an isolated popu-
lation in Germany, the Sorbs [27,28].
At present, about 1000 Sorbs are enrolled in the study.
Sampling comprised unrelated subjects as well as
families. Extensive phenotyping included standardised
questionnaires for past medical history and family his-
tory, collection of anthropometric data (weight, height,
waist-to-hip-ratio, body impedance analysis (BIA)) and a
75g-Glucose-tolerance-test (OGTT). Insulin was mea-
sured with the AutoDELFIA®Insulin assay (PerkinElmer
Life and Analytical Sciences, Turku, Finland). Serum
glucose was measured by the hexokinase method (Auto-
mated analyser Modular, Roche Diagnostics, Mannheim,
Germany). Cases were defined as subjects with gall-
stones in the ultrasonic testing or those who underwent
cholecystectomy due to symptomatic gallstones. Subjects
with ultrasonic exclusion of gallstones were included as
controls in this study. There were no other specific
inclusion and exclusion criteria. A summary sample
description is presented in Table 1. Currently, the entire
family structure of all subjects is not yet known there-
fore, the estimated effect sizes might be biased by cryp-
tic relatedness.
The study was approved by the ethics committee of
the University of Leipzig and all subjects provided
written informed consent before taking part in the
study.
Genotyping of rs9514089
Genotyping of rs9514089 was performed using the Taq-
Man allelic discrimination assay (Assays-on-Demand
(TM), SNP Genotyping Products; Applied Biosystems,
Inc.) on an ABI PRISM 7500 sequence detector (Applied
Biosystems Inc.) according to the manufacturer’s proto-
col. The genotype distribution was consistent with
Hardy-Weinberg equilibrium (minor allele frequency =
35%; p > 0.05). Genotyping success rate was >99%, and
duplicate genotyping concordance was 100%.
Statistics
Standard descriptive and comparative statistics
(ANOVA) were used to characterize and compare clini-
cal parameters in cases and controls Genetic associa-
tions were assessed by linear or logistic regression using
an additive model of inheritance unless stated otherwise,
and adjusted for age, gender and BMI. All effect direc-
tions were standardized to the minor allele.
To obtain the combined effect of the three cohorts we
performed a meta-analysis using the metan command in
STATA based on the estimated effect sizes of each
study and their 95% confidence intervals. The meta-ana-
lysis was performed in a fixed effect model by using the
Mantel-Haenszel method.
All statistical analyses were performed using the SPSS
15.0.1 software package (SPSS, Inc.; Chicago, IL, USA)
and STATA version 9.0, (StataCorp LP, Texas, USA).
Results
Associations of SLC10A2 variant rs9514089 with
gallstones in the Sorbs
In our study, which includes 826 controls and 183
patients with gallstones, rs9514089 did not show any
significant effect on gallstone prevalence, neither in the
additive nor in the recessive or dominant mode of
inheritance (all p > 0.05). In the subgroup of subjects
with BMI ≤ 26 kg/m2 variant rs9514089 tended to be
associated with gallstones (p = 0.05, OR = 0.57), whereas
there was no effect in the group with BMI > 26 kg/m2
(p = 0.52) (Table 2). In the subgroup of females with
Table 1 Main characteristics of the Sorbs
cases controls
N mean ± SD N mean ± SD p value ANOVA
183 (139 f/44 m) 826 (465 f/361 m) < 0.001
age [years] 183 60.5 ± 11.7 826 45.0 ± 15.8 < 0.001
BMI [kg/m2] 180 30.03 ± 5.6 818 26.21 ± 4.5 < 0.001
fat mass [%] 180 26.9 ± 10.4 818 19.95 ± 8.5 < 0.001
waist circumference [cm] 183 98.6 ± 13.3 826 89.1 ± 13.5 < 0.001
hip circumference [cm] 183 109.1 ± 10.5 825 103.7 ± 32.3 0.025
Arithmetic means and standard deviation (SD) are presented for cases, and controls. Univariate ANOVA was performed to assess group differences between
gallstone carriers (cases) and controls.
Tönjes et al. BMC Medical Genetics 2011, 12:149
http://www.biomedcentral.com/1471-2350/12/149
Page 2 of 6
BMI ≤ 26 kg/m2 the effects on gallstone risk reached
nominal level of significance (p = 0.045, OR = 0.51, 188
controls, 102 cases, N(GG carriers) = 38, N(GA carriers) =
124, N(AA carriers) = 131), but would not sustain correc-
tion for multiple testing.
Association with extended phenotypes in the Sorbs
There was no significant association of rs9514089 with
serum parameters of lipid metabolism (p = 0.17 for total
cholesterol, p = 0.78 for HDL-cholesterol, p = 0.10 for
LDL-cholesterol, p = 0.51 for triglycerides, p = 0.17 for
APO-B) in the subgroup of Sorbs without lipid-lowering
medication (Table 3).
Rs9514089 did not show any effects on BMI or fat
mass, neither in the total cohort, nor in females or
males separately. In a subgroup of subjects with avail-
able birth weight data (285 cases, 35 controls), the SNP
tended to be associated with birth weight (p = 0.03).
Gallstone carriers were characterized by increased age,
BMI, fat mass, waist- and hip circumference (Table 1).
In addition, women were affected more frequently than
men.
Meta-analysis
In a combined analysis of the Sorbs and two previously
published cohorts from Aachen and Stuttgart [19] there
was no significant effect of the rs9514089 genotype on
gallstone risk (Figure 1). Also the stratified analyses and
assessment of different genetic models did not reveal
any consistent significant effect on gallstone risk (Table
4). Furthermore, rs9514089 did not show any significant
effect on serum cholesterol levels in the combined ana-
lysis of the three cohorts (p = 0.91, beta 0.001 [-0.014;
0.016] in the additive model) (Table 5).
Table 2 Association of rs9514089 with gallstones in the Sorbs
Trait N p (ADD Mode) OR (ADD) p (dom) OR (dom) p (rec) OR (rec)
(cases/controls) [95% CI] [95% CI] [95% CI]
gallstone (yes versus no) all 183/826 0.19 0.83 0.47 0.87 0.10 0.60
[0.63;1.09] [0.61;1.26] [0.32;1.11]
gallstone (yes versus no) BMI ≤ 26 49/435 0.05 0.57 0.14 0.58 0.09 0.30
[0.33;1.01] [0.29;1.19] [0.07;1.21]
gallstone (yes versus no) BMI > 26 134/391 0.52 0.92 0.85 0.97 0.29 0.72
[0.66;1.26] [0.63;1.49] [0.36;1.47]
gallstone (yes versus no) female 139/465 0.11 0.77 0.19 0.75 0.18 0.62
[0.55;1.06] [0.48;1.16] [0.31:1.26]
gallstone (yes versus no) female and BMI ≤ 26 39/275 0.05 0.51 0.14 0.54 0.07 0.20
[0.26;0.99] [0.24;1.22] [0.33;1.19]
gallstone (yes versus no) male 44/361 0.95 1.02 0.49 1.28 0.34 0.53
[0.61;1.70] [0.64;2.55] [0.15;1.95]
gallstone (OP versus no) all 114/826 0.51 0.89 0.90 0.97 0.25 0.64
[0.64;1.24] [0.62;1.52] [0.31;1.35]
Association of rs9514089 with gallstone risk. Analyses were performed in the additive (ADD), dominant (dom) and recessive (rec) mode of inheritance by logistic
regression analyses. In the total cohort age, gender and BMI were used as covariates, in the stratified analyses/subgroup analyses only age and gender or BMI,
respectively. Results are presented as odds ratios (OR) and their 95% confidence intervals (95% CI). Effect directions were standardized to the minor allele.
Table 3 Association of rs9514089 with metabolic traits in the Sorbs
Trait N p (ADD Mode) beta (ADD) SE beta
total cholesterol [mmol/l] 907 0.17 0.011 0.008
HDL cholesterol [mmol/l] 907 0.78 0.003 0.010
LDL cholesterol [mmol/l] 907 0.10 0.021 0.013
Triglycerides [mmol/l] 907 0.41 0.018 0.022
Apolipoprotein-B [g/l] 905 0.17 0.015 0.011
fasting insulin [pmol/l] 886 0.42 -0.020 0.025
fasting glucose [mmol/l] 886 0.46 0.003 0.004
BMI [kg/m2] 1019 0.23 -0.009 0.007
fat mass [%] 1019 0.91 0.002 0.017
Association of rs9514089 with components of the metabolic syndrome. Analyses were performed in the additive (ADD) mode of inheritance by linear regression
analyses. Age, gender and BMI were used as covariates (for BMI and fat mass only age and gender). Results are presented as betas and their standard errors (SE).
Effect directions were standardized to the minor allele. For analyses of glucose and insulin only non-diabetic subjects were included, effects on lipid traits were
only assessed in subjects without lipid-lowering medication.
Tönjes et al. BMC Medical Genetics 2011, 12:149
http://www.biomedcentral.com/1471-2350/12/149
Page 3 of 6
Discussion
Recently, SLC10A2 was suggested as a novel susceptibil-
ity gene for cholelithiasis in humans [19]. The data by
Renner et al. indicated that the SNP effects were most
pronounced when calculated in the recessive mode of
inheritance, i.e. the risk of gallstone development was
highest in subjects homozygous for the rs9514089 G
allele. Therefore, we performed a meta-analysis includ-
ing three independent cohorts (previously published
cohorts from Stuttgart and Aachen and the cohort of
Sorbs) and assuming a recessive mode of inheritance.
Although there was no significant difference in minor
allele frequencies between the three cohorts the risk
allele was the G allele in the Stuttgart and Aachen
cohorts but the A allele in the Sorbs. This effect is most
likely driven by the small samples sizes in each genotype
group in the cohorts. Furthermore, gene-environmental
interactions based on factors such as eating behavior
and nutrition composition as well as gene-gene interac-
tions should be taken into account. In our study the
meta-analysis of all three cohorts did not indicate any
significant effect of the rs9514089 genotype on gallstone
risk. The Sorbs and the Aachen cohort showed a trend
of association in the subgroup of non-obese individuals
indicating a possible relationship between the SNP effect
and obesity as suggested by Renner et al. [19]. This
would be in line with the reported weight specific regu-
lation of the ASBT expression. However, so far, a dimin-
ished ASBT expression was confirmed only in non-
obese gallstone patients [25,26]. Also, since effect direc-
tion of the rs9514089 variant is different in the Sorbs
the meta-analysis did not reveal any significant associa-
tion in the combined analysis of all three cohorts.
In the Sorbs a nominal association of the SLC10A2
genotype and cholelithiasis was observed in the non-
obese subgroup, even though the SNP effect could only
be detected in women. This may be due to skewed gen-
der distribution in favor of women and thereby greater
statistical power for the female subset. The differences
between the populations do not necessarily represent
population specific effects of the gene variant. Consis-
tent with the effect in the entire cohort the effect direc-
tion also in this subgroup was opposite in the Sorbs
compared with the Aachen cohort. This could be attrib-
uted to population specific environmental factors or a
Table 4 Meta-analysis of effects of rs9514089 on gallstone risk in the Sorbs, Aachen cohort and Stuttgart cohort
Sorbs Aachen Stuttgart
Trait p (rec) OR (rec) p (rec) OR (rec) p (rec) OR (rec) p meta OR meta
[95% CI] [95% CI] [95% CI] [95% CI]
gallstone (yes versus no) all 0.10 0.60 0.01 2.19 0.46 1.5 0.41 1.19
[0.32;1.11] [1.19;4.17] [0.51;4.43] [0.79; 1.78]
gallstone (yes versus no) BMI ≤ 26 0.09 0.30 0.04 2.59 0.11 4.0 0.24 1.54
[0.07;1.21] [1.02;7.01] [0.67;23.95] [0.75; 3.17]
gallstone (yes versus no) BMI > 26 0.29 0.72 0.12 1.88 0.65 0.72 0.99 0.99
[0.36;1.47] [0.80;4.62] [0.18;2.96] [0.60; 1.66]
gallstone (yes versus no) female 0.18 0.62 0.18 1.81 0.99 0.99 0.79 1.08
[0.31;1.26] [0.78;4.33] [0.27;3.58] [0.61; 1.90]
gallstone (yes versus no) male 0.34 0.53 0.03 2.79 0.37 2.5 0.22 1.59
[0.15;1.95] [1.07;7.92] [0.32;19.54] [0.76; 3.32]
Meta-analysis of three independent cohorts (Sorbs - 183 cases/826 controls, Stuttgart - 56 cases/71 controls, Aachen - 184 cases/184). Analyses were performed
under the recessive mode of inheritance in a fixed effects model. Results are presented as odds ratios (OR) and their 95% confidence intervals (95% CI). Effect
directions were standardized to the minor allele.
Stuttgart
Aachen
Sorbs
male
Stuttgart
Aachen
Sorbs
female
Stuttgart
Aachen
Sorbs
obese
Stuttgart
Aachen
Sorbs
lean
Stuttgart
Aachen
Sorbs
all
OR
19.5
24
Figure 1 Meta-analysis of effects of rs9514089 on gallstone
risk in three independent cohorts.
Tönjes et al. BMC Medical Genetics 2011, 12:149
http://www.biomedcentral.com/1471-2350/12/149
Page 4 of 6
different phenotype distribution in each population or it
could be simply due to the lack of power based on the
small samples sizes of the cohorts.
Furthermore, our meta-analysis did not confirm the
reported effects of rs9514089 on serum total cholesterol
[19]. However, the link between circulating lipid levels
and development of gallstones has not been completely
elucidated.
Conclusion
The effect of rs9514089 genotype on gallstone risk was
not replicated in the Sorbs. Further analyses in larger
cohorts are required to finally assess the role of genetic
variants in SLC10A2 in human gallstone development
and lipid metabolism
Acknowledgements
We thank all those who participated in the studies. We also thank Mrs. Ines
Müller and Beate Enigk for excellent technical assistance. This work was
supported by grants from the Interdisciplinary Centre for Clinical Research
Leipzig at the Faculty of Medicine of the University of Leipzig (Project N06
to P.K. and B27 to M.S. P.K. and A.T.), from the German Diabetes Association
(to A.T., P.K.) and the Diabetes Hilfs- und Forschungsfonds Deutschland
DHFD (to M.S., A.T., P.K.). H.W. was supported by the Deutsche
Forschungsgemeinschaft (WI1905 3-2). The study of Renner et al was
supported by the Robert Bosch Foundation (P4-1/03). F.L. was supported by
Saarland University (HOMFOR T201000385).
Author details
1Department of Medicine, Division of Endocrinology and Nephrology,
University of Leipzig, Leipzig, Germany. 2Department of Medicine, Division of
Gastroenterology and Rheumatology, University of Leipzig, Leipzig, Germany.
3Dr. Margarete Fischer-Bosch Institute of Clinical Pharmacology and
University of Tuebingen, Stuttgart, Germany. 4Department of Internal
Medicine I, Robert Bosch Hospital, Stuttgart, Germany. 5Department of
Internal Medicine II, University Saarland, Homburg, Germany.
6Interdisciplinary Centre for Clinical Research, University of Leipzig, Leipzig,
Germany.
Authors’ contributions
AT performed the phenotyping in the Sorbs, carried out the statistical
analysis and drafted the manuscript. OR, SH, FL and EFS performed the
genotyping and statistical analyses in the cohorts from Aachen and
Stuttgart. PK and JH carried out the molecular genetic studies in the Sorbs.
FL and HW participated in the design of the study. AT, PK and MS
conceived the study, participated in its design and coordination and helped
to draft the manuscript. All authors read and approved the final manuscript.
Competing interests
The authors declare that they have no competing interests.
Received: 8 September 2010 Accepted: 17 November 2011
Published: 17 November 2011
References
1. Hou L, Shu XO, Gao YT, Ji BT, Weiss JM, Yang G, Li HL, Blair A, Zheng W,
Chow WH: Anthropometric measurements, physical activity, and the risk
of symptomatic gallstone disease in Chinese women. Ann Epidemiol
2009, 19:344-50.
2. La Vecchia C, Negri E, D’Avanzo B, Franceschi S, Boyle P: Risk factors for
gallstone disease requiring surgery. Int J Epidemiol 1991, 20:209-15.
3. Bodmer M, Brauchli YB, Krahenbuhl S, Jick SS, Meier CR: Statin use and risk
of gallstone disease followed by cholecystectomy. JAMA 2009,
302:2001-7.
4. Kovacs P, Kress R, Rocha J, Kurtz U, Miquel JF, Nervi F, Mendez-Sanchez N,
Uribe M, Bock HH, Schirin-Sokhan R, Stumvoll M, Mossner J, Lammert F,
Wittenburg H: Variation of the gene encoding the nuclear bile salt
receptor FXR and gallstone susceptibility in mice and humans. J Hepatol
2008, 48:116-24.
5. Dixit M, Choudhuri G, Saxena R, Mittal B: Association of apolipoprotein A1-
C3 gene cluster polymorphisms with gallstone disease. Can J
Gastroenterol 2007, 21:569-75.
6. Grunhage F, Acalovschi M, Tirziu S, Walier M, Wienker TF, Ciocan A,
Mosteanu O, Sauerbruch T, Lammert F: Increased gallstone risk in humans
conferred by common variant of hepatic ATP-binding cassette
transporter for cholesterol. Hepatology 2007, 46:793-801.
7. Xie Y, Newberry EP, Kennedy SM, Luo J, Davidson NO: Increased
susceptibility to diet-induced gallstones in liver fatty acid binding
protein knockout mice. J Lipid Res 2009, 50:977-87.
8. Rudkowska I, Jones PJ: Polymorphisms in ABCG5/G8 transporters linked
to hypercholesterolemia and gallstone disease. Nutr Rev 2008, 66:343-8.
9. Hoblinger A, Lammert F: Genetics of biliary tract diseases: new insights
into gallstone disease and biliary tract cancers. Curr Opin Gastroenterol
2008, 24:363-71.
10. Chang SC, Rashid A, Gao YT, Andreotti G, Shen MC, Wang BS, Han TQ,
Zhang BH, Sakoda LC, Leitzmann MF, Chen BE, Rosenberg PS, Chen J,
Chanock SJ, Hsing AW: Polymorphism of genes related to insulin
sensitivity and the risk of biliary tract cancer and biliary stone: a
population-based case-control study in Shanghai, China. Carcinogenesis
2008, 29:944-8.
11. Lammert F, Miquel JF: Gallstone disease: from genes to evidence-based
therapy. J Hepatol 2008, 48(Suppl 1):S124-S135.
12. Schafmayer C, Volzke H, Buch S, Egberts J, Spille A, von Eberstein H,
Franke A, Seeger M, Hinz S, Elsharawy A, Rosskopf D, Brosch M, Krawczak M,
Foelsch UR, Schafmayer A, Lammert F, Schreiber S, Faendrich F, Hampe J,
Tepel J: Investigation of the Lith6 candidate genes APOBEC1 and PPARG
in human gallstone disease. Liver Int 2007, 27:910-9.
13. Grunhage F, Lammert F: Gallstone disease. Pathogenesis of gallstones: A
genetic perspective. Best Pract Res Clin Gastroenterol 2006, 20:997-1015.
14. Lyons MA, Wittenburg H: Cholesterol gallstone susceptibility loci: a
mouse map, candidate gene evaluation, and guide to human LITH
genes. Gastroenterology 2006, 131:1943-70.
Table 5 Meta-analysis of effects of rs9514089 on serum cholesterol and triglyceride levels in the Sorbs, Aachen cohort
and Stuttgart cohort
Sorbs Aachen Stuttgart
Trait p (ADD) beta (ADD) p (ADD) beta (ADD) p (ADD) beta (ADD) p meta beta meta
[95% CI] [95% CI] [95% CI] [95% CI]
total cholesterol [mmol/l] 0.17 0.01 0.28 -0.024 0.06 -0.048 0.91 0.001
[-0.005; 0.028] [-0.067; 0.019] [-0.097; 0.001] [-0.014; 0.016]
triglycerides [mmol/l] 0.41 0.018 0.92 -0.004 0.37 -0.062 0.69 0.007
[-0.025; 0.061] [-0.078; 0.070] [-0.197; 0.073] [-0.029; 0.043]
Meta-analysis of three independent cohorts (Sorbs N = 907, Aachen N = 358, Stuttgart N = 122). Analyses were performed under the additive mode of
inheritance in a fixed effects model. Only subjects without any lipid-lowering medication were included in the analyses. Results were assessed in a linear
regression model adjusted for age, sex and BMI and effect directions are standardized to the minor allele.
Tönjes et al. BMC Medical Genetics 2011, 12:149
http://www.biomedcentral.com/1471-2350/12/149
Page 5 of 6
