Sibling number and prevalence of allergic disorders in pregnant Japanese women: baseline data from the Kyushu Okinawa Maternal and Child Health Study
© Miyake et al; licensee BioMed Central Ltd. 2011
Received: 23 October 2010
Accepted: 14 July 2011
Published: 14 July 2011
Although an inverse relationship between number of siblings and likelihood of allergic disorders has been shown in many epidemiological studies, the biological mechanism underlying this phenomenon has not yet been identified. There is no epidemiological research regarding the sibling effect on allergic disorders in Japanese adults. The current cross-sectional study examined the relationship between number of siblings and prevalence of allergic disorders among adult women in Japan.
Subjects were 1745 pregnant women. This study was based on questionnaire data. The definitions of wheeze and asthma were based on criteria from the European Community Respiratory Health Survey whereas those of eczema and rhinoconjunctivitis were based on criteria from the International Study of Asthma and Allergies in Childhood. Adjustment was made for age, region of residence, pack-years of smoking, secondhand smoke exposure at home and at work, family history of asthma, atopic eczema, and allergic rhinitis, household income, and education.
The prevalence values of wheeze, asthma, eczema, and rhinoconjunctivitis in the past 12 months were 10.4%, 5.5%, 13.0%, and 25.9%, respectively. A significant inverse exposure-response relationship was observed between the number of older siblings and rhinoconjunctivitis, but not wheeze, asthma, or eczema (P for trend = 0.03); however, the adjusted odds ratio (OR) for having 2 or more older siblings was not significant although the adjusted OR for having 1 older sibling was statistically significant (adjusted OR = 0.71 [95% CI: 0.56-0.91]). Number of total siblings and number of younger siblings were not related to wheeze, asthma, eczema, or rhinoconjunctivitis.
This study found a significant inverse relationship between the number of older siblings and the prevalence of rhinoconjunctivitis among pregnant Japanese women. Our findings are likely to support the intrauterine programming hypothesis; however, we could not rule out the hygiene hypothesis.
Although an inverse relationship between number of siblings and likelihood of allergic disorders has been shown in many epidemiological studies, the biological mechanism underlying this phenomenon has not yet been identified . Strachan first proposed the hygiene hypothesis in 1989, suggesting that exposure to infections transmitted by unhygienic contact with older or younger siblings might influence the skewing of the Th1/Th2 balance away from allergy-promoting Th2 cells toward Th1 cells, leading to a reduced risk of allergic disorders [2, 3]. However, a study of young adults in the UK showed that both the number of gastrointestinal infections before the age of 5 years and the number of brothers were independently inversely related to the prevalence of atopy after mutual adjustment for the 2 factors .
The intrauterine programming hypothesis has been proposed as an alternative explanation of the mechanism of the sibling effect . According to this theory, maternal immunomodulation acquired through multiple pregnancies might be transmitted to the developing fetus [5–7]. Although a cohort study in the USA found no association between birth order and total cord blood IgE level , another recent study reported a significant interaction between birth order and IL13 polymorphisms affecting allergic sensitization: the effect of IL13 was restricted only to firstborn children, suggesting adverse prenatal programming in firstborn offspring .
To our knowledge, there is no evidence that number of siblings affects the prevalence of allergic disorders among Japanese adults, though such evidence does exist for Japanese schoolchildren, as shown in our previous studies [10, 11]. The current cross-sectional study examined the relationship between number of siblings and the prevalence of allergic disorders in Japanese pregnant women using baseline data from the Kyushu Okinawa Maternal and Child Health Study (KOMCHS).
The KOMCHS is an ongoing prospective prebirth cohort study that investigates risk and preventive factors for maternal and child health problems such as allergic disorders. The KOMCHS requested that 131 obstetric hospitals in Fukuoka Prefecture, the largest prefecture on Kyushu Island in southern Japan, with a total population of approximately 5.04 million, provide as many pregnant women as possible with a set of leaflets explaining the KOMCHS, an application form to participate in the study, and a self-addressed and stamped return envelope during the period from April 2007 to March 2008. From May 2007 to March 2008, the KOMCHS also requested that 40 obstetric hospitals in Okinawa Prefecture, an island in the southernmost area of Japan, with a total population of almost 1.37 million, provide pregnant women with the same documents. In addition, to increase the sample size, from August 2007 to March 2008, pregnant women living in 6 prefectures on Kyushu Island other than Fukuoka Prefecture, with a total population of approximately 8.22 million, were provided with these documents at 252 obstetric hospitals. Pregnant women who intended to participate in the KOMCHS returned the application form to the data management center. In the end, a total of 1757 pregnant women between the 5th and 39th week of pregnancy gave their fully informed consent in writing to participate and completed the baseline survey. Excluded were 12 pregnant women because of missing data on the factors under study, leaving data on 1745 pregnant women available for analysis. The ethics committee of the Faculty of Medicine, Fukuoka University approved the KOMCHS.
In the baseline survey, each participant filled out a set of 2 self-administered questionnaires. Participants mailed the answered questionnaires to the data management center. Research technicians completed missing or illogical data by telephone interview.
One of the self-administered questionnaires included questions on wheeze and asthma based on the European Community Respiratory Health Survey  and questions on eczema and rhinoconjunctivitis based on the International Study of Asthma and Allergies in Childhood [13, 14]. Wheeze was defined as a positive response to the question, 'Have you had wheezing or whistling in your chest at any time in the last 12 months?' Asthma was defined as either having experienced an asthma attack during the last 12 months or currently using asthma medication. Affirmative answers to the following 3 questions were required to indicate the presence of eczema: 'Have you ever had an itchy rash which was coming and going for at least 6 months?', 'Have you had this itchy rash at any time in the last 12 months?' and 'Has this itchy rash at any time affected any of the following places: the folds of the elbows, behind the knees, in front of the ankles, under the buttocks, or around the neck, ears, or eyes?' Rhinoconjunctivitis was defined as a positive response to the following 2 questions: 'In the last 12 months, have you had a problem with sneezing or a runny or blocked nose when you did not have a cold or flu?' and 'In the last 12 months, has this nose problem been accompanied by itchy-watery eyes?' The questionnaire also elicited information on age, number of older and younger siblings, smoking habits, secondhand smoke exposure at home and at work, family history of asthma, atopic eczema, and allergic rhinitis, household income, and education. A family history of asthma, atopic eczema, or allergic rhinitis (including Japanese cedar pollinosis) was considered to be present if one or more parents or siblings of the study subjects had been diagnosed by a physician with any of these allergic disorders.
The second questionnaire was a validated self-administered diet history questionnaire. Data regarding diet were not used in the present study.
Age, region of residence, pack-years of smoking, secondhand smoke exposure at home and at work, family history of asthma, atopic eczema, and allergic rhinitis, household income, and education were a priori selected as potential confounding factors. Our previous study using baseline data from the Osaka Maternal and Child Health Study showed that current smoking was significantly associated with a higher prevalence of asthma whereas secondhand smoke exposure was positively significantly related to the prevalence of allergic rhinitis . Region of residence was classified into 3 categories (Fukuoka Prefecture, other than Fukuoka Prefecture on Kyushu Island, and Okinawa Prefecture), pack-years of smoking into 3 (none, 0.05-3.9, and ≥ 4.0), secondhand smoke exposure at home and at work into 2 (never and ever), household income into 3 (< 4,000,000, 4,000,000-5,999,999, and ≥ 6,000,000 yen/year), education into 3 (< 13, 13-14, and ≥ 15 years), total number of siblings into 4 (0, 1, 2, and ≥ 3), number of older siblings into 3 (0, 1, and ≥ 2) and number of younger siblings into 3 (0, 1, and ≥ 2). Age was used as a continuous variable.
Multiple logistic regression analysis was performed to estimate adjusted odds ratios (ORs) and 95% confidence intervals (CIs) of allergic disorders relative to the number of siblings. Trend of association was assessed by a logistic regression model assigning consecutive integers to the categories of the exposure variables. All statistical analyses were performed using the SAS software package version 9.2 (SAS Institute, Inc., Cary, NC, USA).
Distribution of selected characteristics in 1745 pregnant women, Kyushu Okinawa Maternal and Child Health Study, Japan
Region of residence
Other than Fukuoka Prefecture in Kyushu
Pack-years of smoking
Ever secondhand smoke exposure at home
Ever secondhand smoke exposure at work
Family history of asthma
Family history of atopic eczema
Family history of allergic rhinitis
Household income (yen/year)
Total no. siblings
No. older siblings
No. younger siblings
Adjusted odds ratios (ORs) and 95% confidence intervals (CIs) for allergic disorders according to the number of siblings in 1745 pregnant women, Kyushu Okinawa Maternal and Child Health Study, Japan1
OR (95% CI)
OR (95% CI)
OR (95% CI)
OR (95% CI)
0 (n = 94)
1 (n = 788)
2 (n = 688)
≥ 3 (n = 175)
P for trend
0 (n = 829)
1 (n = 622)
≥ 2 (n = 294)
P for trend
0 (n = 695)
1 (n = 695)
≥ 2 (n = 355)
P for trend
When we stratified our subjects into two groups based on the presence or absence of older siblings, subjects with 1 or more older siblings had a significantly lower prevalence of rhinoconjunctivitis than subjects without older siblings did, but there was no such association between the presence of older siblings and wheeze, asthma, or eczema: the adjusted OR was 0.73 (95% CI: 0.59-0.91). Neither the presence of younger siblings nor the presence of siblings of any age was related to any of the outcomes under study.
The current study found a significant inverse exposure-response relationship between the number of older siblings and the prevalence of rhinoconjunctivitis, but not wheeze, asthma, or eczema. Total sibship size and number of younger siblings were not materially associated with the prevalence of wheeze, asthma, eczema, or rhinoconjunctivitis.
Epidemiological evidence for the sibling effect on allergic disorders among adults is limited. A cross-sectional study of men and women in the UK aged 16-30 years found significant inverse exposure-response associations of sibship size and birth order with the prevalence of hay fever whereas larger sibship size, but not birth order, was significantly inversely related to the prevalence of eczema-urticaria . In a cross-sectional study of pregnant women in Denmark, larger sibship size was significantly associated with a lower prevalence of allergic rhinitis and asthma with allergic rhinitis, but not asthma without allergic rhinitis; the inverse associations between having older siblings and allergic rhinitis and asthma with allergic rhinitis were more pronounced than the inverse associations between having younger siblings and these disorders . Having 4 or more older siblings was significantly related to a lower prevalence of allergic rhinitis and/or allergic conjunctivitis, but not asthma or atopic dermatitis, over the lifetime of the subjects in a cross-sectional study of Finnish young adults aged 18-24 years . In a retrospective cohort study of Swedish male conscripts, number of older siblings was significantly inversely associated with the prevalence of allergic rhinitis, but not asthma .
Turning to the evidence concerning Japanese children, having 2 or more older siblings was significantly related to a lower prevalence of rhinoconjunctivitis, but not wheeze or eczema, in a cross-sectional study of children aged 12-15 years . Another cross-sectional study of Japanese children aged 6-15 years reported significant exposure-response relationships between larger sibship size and the prevalences of wheeze, asthma, eczema, and rhinoconjunctivitis, and demonstrated that having older siblings reduces the prevalences of eczema and rhinoconjunctivitis while having younger siblings reduces only the prevalence of eczema . These findings, which demonstrate an inverse association between the number of older siblings and allergic rhinitis and null associations between sibship size and wheeze, asthma, or eczema, are in agreement with the current results.
Karmaus and colleagues have found that the level of cord blood IgE was reduced with increasing birth order and that elevated cord blood IgE values were significantly related to an increased risk of allergic sensitization at 48 months of age, though there was no association between birth order and allergic sensitization . Karmaus and colleagues have also shown that maternal IgE levels are significantly reduced with increasing numbers of live offspring . In a study in the UK, women with more pregnancies over a seven-year period were more likely to have lost atopy and symptoms of hay fever, but not asthma, during that period . The fact that maternal immune tolerance against allergens increases as more live offspring are born, and the fact that maternal immune status is communicated to the fetus, might explain the inverse association between the number of older siblings and rhinoconjunctivitis in the current study. Another version of the intrauterine programming hypothesis suggests a potential role of hormones; progesterone and testosterone seem to promote the preferential development of Th2-like cells . In addition, the potential role of endocrine disruptors should be discussed. Parity was significantly associated with decreasing levels of organochlorines in breast milk of Norwegian women . Higher levels of cord blood dichlorodiphenyldichloroethylene were significantly related to an increased risk of asthma at 6.5 years in a cohort study in Spain . Nevertheless, a UK study cited above showed a significant inverse association between sibship size and the prevalence of combined allergies, asthma, eczema-urticaria, and hay fever among people born between 1918 and 1930 when organochlorines had not come into widespread use .
The absence of a sibling effect on wheeze, asthma, or eczema in this study might be ascribed to the attenuation of the sibling effect with advancing age. In our previous study, larger sibship size was significantly associated with the prevalences of wheeze, asthma, eczema, and rhinoconjunctivitis among children aged 6-10 years, but only with the prevalence of eczema among subjects aged 11-15 years . It should be noted that the results of that study, specifically, the findings concerning eczema and rhinoconjunctivitis in the older age group, are at variance with the present findings.
This study had certain methodological advantages. Study subjects were homogeneous in that they were all pregnant. Definitions of wheeze and asthma were based on the questions in the European Community Respiratory Health Survey. Likewise, definitions of eczema and rhinoconjunctivitis were based on the questions in the International Study of Asthma and Allergies in Childhood, although validation tests of such questions have not been performed for young Japanese adults. Several potential confounders were adjusted for, though we could not control for breastfeeding, early day care center attendance, and external factors as aeroallergens and air pollution.
This study also had certain limitations. The participation rate could not be calculated because the exact number of eligible pregnant women who were provided with the KOMCHS documents and application form by the 423 collaborating obstetric hospitals is not available. Nevertheless, the participation rate must have been fairly low, given that the present study used data from only 970 pregnant women who lived in Fukuoka Prefecture, while, according to the government of Fukuoka Prefecture, the number of childbirths was 46,393 in 2007 and 46,695 in 2008. We were not able to assess differences between participants and non-participants because information on personal characteristics such as age, socioeconomic status, and history of allergic disorders was not available for non-participants. Our subjects were probably not a representative sample of Japanese women in the general population. For example, educational levels were higher in the current study population than in the general population. According to the 2000 population census of Japan, the proportions of women aged 30 to 34 years in Fukuoka Prefecture with years of education of < 13, 13-14, ≥ 15, and unknown were 52.0%, 31.5%, 11.8%, and 4.8%, respectively . The corresponding figures for the current study were 24.6%, 33.0%, 42.4%, and 0.0%, respectively. The present population therefore might have had greater awareness about health issues compared to the general population. The lack of significant inverse relationships between total sibship size and the prevalences of wheeze and eczema might be attributable to an insufficient statistical power.
The interface between allergy/immunology and pregnancy should be discussed, as it may have an influence on the association of interest. It has been suggested that pregnancy involves a shift to the Th2 side of the immune response  although Chaouat et al. have pointed out the importance of the role of natural killer cells and IL-12, IL-15, and IL-18 tripods in successful or failed pregnancies in humans beyond the Th1/Th2 paradigm . The hormonal changes in pregnancy are often invoked to explain the apparent association between rhinitis symptoms and pregnancy. However, rhinitis ascribed solely to pregnancy may not be a distinct entity because most pregnant women do not have significant nasal symptoms .
The current study among Japanese women found a significant inverse relationship between the number of older, but not total or younger, siblings and the prevalence of rhinoconjunctivitis, whereas no such sibling effect was observed for wheeze, asthma, and eczema. Our findings are likely to support the intrauterine programming hypothesis, though we could not rule out the hygiene hypothesis.
The authors would like to acknowledge the Kyushu Branch of Japan Allergy Foundation, the Fukuoka Association of Obstetricians & Gynecologists, the Okinawa Association of Obstetricians & Gynecologists, the Miyazaki Association of Obstetricians & Gynecologists, the Oita Association of Obstetricians & Gynecologists, the Kumamoto Association of Obstetricians & Gynecologists, the Nagasaki Association of Obstetricians & Gynecologists, the Kagoshima Association of Obstetricians & Gynecologists, the Saga Association of Obstetricians & Gynecologists, the Fukuoka Society of Obstetrics and Gynecology, the Okinawa Society of Obstetrics and Gynecology, the Fukuoka City Government, and the Fukuoka City Medical Association for their valuable support.
This study was supported by KAKENHI (19590606, 20791654, 21590673, and 22592355) and Health and Labour Sciences Research Grants, Research on Allergic Disease and Immunology from the Ministry of Health, Labour, and Welfare, Japan.
- Karmaus W, Botezan C: Does a higher number of siblings protect against the development of allergy and asthma? A review. J Epidemiol Community Health. 2002, 56: 209-217. 10.1136/jech.56.3.209.View ArticlePubMedPubMed CentralGoogle Scholar
- Strachan DP: Hay fever, hygiene, and household size. BMJ. 1989, 299: 1259-1260. 10.1136/bmj.299.6710.1259.View ArticlePubMedPubMed CentralGoogle Scholar
- Schaub B, Lauener R, von Mutius E: The many faces of the hygiene hypothesis. J Allergy Clin Immunol. 2006, 117: 969-977. 10.1016/j.jaci.2006.03.003.View ArticlePubMedGoogle Scholar
- Cullinan P, Harris JM, Newman Taylor AJ, Jones M, Taylor P, Dave JR, Mills P, Moffat SA, White CW, Figg JK, Moon AM, Barnes MC: Can early infection explain the sibling effect in adult atopy?. Eur Respir J. 2003, 22: 956-961. 10.1183/09031936.03.00039102.View ArticlePubMedGoogle Scholar
- Karmaus W, Arshad H, Mattes J: Does the sibling effect have its origin in utero? Investigating birth order, cord blood immunoglobulin E concentration, and allergic sensitization at age 4 years. Am J Epidemiol. 2001, 154: 909-915. 10.1093/aje/154.10.909.View ArticlePubMedGoogle Scholar
- Karmaus W, Arshad SH, Sadeghnejad A, Twiselton R: Does maternal immunoglobulin E decrease with increasing order of live offspring? Investigation into maternal immune tolerance. Clin Exp Allergy. 2004, 34: 853-859. 10.1111/j.1365-2222.2004.01959.x.View ArticlePubMedGoogle Scholar
- Harris JM, White C, Moffat S, Mills P, Newman Taylor AJ, Cullinan P: New pregnancies and loss of allergy. Clin Exp Allergy. 2004, 34: 369-372. 10.1111/j.1365-2222.2004.01905.x.View ArticlePubMedGoogle Scholar
- Wegienka G, Havstad S, Shue L, Zoratti E, Ownby DR, Johnson CC: Birth order and cord immunoglobulin E: results using a high-sensitivity immunoglobulin E protocol. Int Arch Allergy Immunol. 2008, 145: 305-312. 10.1159/000110889.View ArticlePubMedGoogle Scholar
- Ogbuanu IU, Karmaus WJ, Zhang H, Sabo-Attwood T, Ewart S, Roberts G, Arshad SH: Birth order modifies the effect of IL13 gene polymorphisms on serum IgE at age 10 and skin prick test at ages 4, 10 and 18: a prospective birth cohort study. Allergy Asthma Clin Immunol. 2010, 6: 6-10.1186/1710-1492-6-6.View ArticlePubMedPubMed CentralGoogle Scholar
- Miyake Y, Yura A, Iki M: Cross-sectional study of allergic disorders in relation to familial factors in Japanese adolescents. Acta Paediatr. 2004, 93: 380-385.View ArticlePubMedGoogle Scholar
- Ohfuji S, Miyake Y, Arakawa M, Tanaka K, Sasaki S: Sibship size and prevalence of allergic disorders in Japan: The Ryukyus Child Health Study. Pediatr Allergy Immunol. 2009, 20: 377-384. 10.1111/j.1399-3038.2008.00804.x.View ArticlePubMedGoogle Scholar
- European Community Respiratory Health Survey: Variations in the prevalence of respiratory symptoms, self-reported asthma attacks, and use of asthma medication in the European Community Respiratory Health Survey (ECRHS). Eur Respir J. 1996, 9: 687-695.View ArticleGoogle Scholar
- Williams H, Robertson C, Stewart A, Aït-Khaled N, Anabwani G, Anderson R, Asher I, Beasley R, Björkstén B, Burr M, Clayton T, Crane J, Ellwood P, Keil U, Lai C, Mallol J, Martinez F, Mitchell E, Montefort S, Pearce N, Shah J, Sibbald B, Strachan D, von Mutius E, Weiland SK: Worldwide variations in the prevalence of symptoms of atopic eczema in the international study of asthma and allergies in childhood. J Allergy Clin Immunol. 1999, 103: 125-138. 10.1016/S0091-6749(99)70536-1.View ArticlePubMedGoogle Scholar
- Strachan D, Sibbald B, Weiland S, Aït-Khaled N, Anabwani G, Anderson HR, Asher MI, Beasley R, Björkstén B, Burr M, Clayton T, Crane J, Ellwood P, Keil U, Lai C, Mallol J, Martinez F, Mitchell E, Montefort S, Pearce N, Robertson C, Shah J, Stewart A, von Mutius E, Williams H: Worldwide variations in prevalence of symptoms of allergic rhinoconjunctivitis in children: the International Study of Asthma and Allergies in Childhood (ISAAC). Pediatr Allergy Immunol. 1997, 8: 161-176. 10.1111/j.1399-3038.1997.tb00156.x.View ArticlePubMedGoogle Scholar
- Miyake Y, Miyamoto S, Ohya Y, Sasaki S, Matsunaga I, Yoshida T, Hirota Y, Oda H, The Osaka Maternal and Child Health Study Group: Association of active and passive smoking with allergic disorders in Japanese pregnant females: baseline data from the Osaka Maternal and Child Health Study. Ann Allergy Asthma Immunol. 2005, 94: 644-651. 10.1016/S1081-1206(10)61322-1.View ArticlePubMedGoogle Scholar
- Kinra S, Davey Smith G, Jeffreys M, Gunnell D, Galobardes B, McCarron P: Association between sibship size and allergic diseases in the Glasgow Alumni Study. Thorax. 2006, 61: 48-53.View ArticlePubMedPubMed CentralGoogle Scholar
- Westergaard T, Rostgaard K, Wohlfahrt J, Andersen PK, Aaby P, Melbye M: Sibship characteristics and risk of allergic rhinitis and asthma. Am J Epidemiol. 2005, 162: 125-132. 10.1093/aje/kwi169.View ArticlePubMedGoogle Scholar
- Kilpeläinen M, Terho EO, Helenius H, Koskenvuo M: Farm environment in childhood prevents the development of allergies. Clin Exp Allergy. 2000, 30: 201-208. 10.1046/j.1365-2222.2000.00800.x.View ArticlePubMedGoogle Scholar
- Bråbäck L, Hedberg A: Perinatal risk factors for atopic disease in conscripts. Clin Exp Allergy. 1998, 28: 936-942. 10.1046/j.1365-2222.1998.00282.x.View ArticlePubMedGoogle Scholar
- Polder A, Skaare JU, Skjerve E, Løken KB, Eggesbø M: Levels of chlorinated pesticides and polychlorinated biphenyls in Norwegian breast milk (2002-2006), and factors that may predict the level of contamination. Sci Total Environ. 2009, 407: 4584-4590. 10.1016/j.scitotenv.2009.04.032.View ArticlePubMedGoogle Scholar
- Sunyer J, Torrent M, Garcia-Esteban R, Ribas-Fitó N, Carrizo D, Romieu I, Antó JM, Grimalt JO: Early exposure to dichlorodiphenyldichloroethylene, breastfeeding and asthma at age six. Clin Exp Allergy. 2006, 36: 1236-1241. 10.1111/j.1365-2222.2006.02560.x.View ArticlePubMedGoogle Scholar
- Statistics Bureau, Ministry of Public Management, Home Affairs, Posts and Telecommunications: 2000 Population Census of Japan, Vol. 3-2-40, Labour Force Status of Population, Industry (Major Groups) of Employed Persons, and Education: Fukuoka-ken. 2002, Tokyo: Statistics Bureau, Ministry of Public Management, Home Affairs, Posts and Telecommunications, (in Japanese)Google Scholar
- Palmer GW, Claman HN: Pregnancy and immunology: selected aspects. Ann Allergy Asthma Immunol. 2002, 89: 350-359. 10.1016/S1081-1206(10)62034-0.View ArticlePubMedGoogle Scholar
- Chaouat G, Ledée-Bataille N, Dubanchet S, Zourbas S, Sandra O, Martal J: TH1/TH2 paradigm in pregnancy: paradigm lost? Cytokines in pregnancy/early abortion: reexamining the TH1/TH2 paradigm. Int Arch Allergy Immunol. 2004, 134: 93-119. 10.1159/000074300.View ArticlePubMedGoogle Scholar
- The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1471-2458/11/561/prepub