HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Supplemental data Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Google Scholar Articles by Timpson, N. J Articles by Davey Smith, G. PubMed PubMed Citation Articles by Timpson, N. J Articles by Davey Smith, G. Agricola Articles by Timpson, N. J Articles by Davey Smith, G.

The fat mass–and obesity-associated locus and dietary intake in children^1,2,3

¹ MRC CAiTE Centre, Department of Social Medicine (NJT and GDS) and the Department of Community Based Medicine (PME), Bristol University, Bristol, United Kingdom; Genetics of Complex Traits (TMF and ATH) and Diabetes Genetics (TMF and ATH), Institute of Biomedical and Clinical Science, Peninsula Medical School, Exeter, United Kingdom; the School of Pharmacy and Biomolecular Sciences, University of Brighton, Brighton, United Kingdom (IR); the Oxford Centre for Diabetes, Endocrinology and Metabolism (MIM) and the Wellcome Trust Centre for Human Genetics (NJT and MIM), University of Oxford, Oxford, United Kingdom

² Supported by The UK Medical Research Council, the Wellcome Trust, and the University of Bristol.

³ Reprints not available. Address correspondence to NJ Timpson, Department of Social Medicine, Bristol University, Canynge Hall, Whiteladies Road, Bristol, BS8 2PR United Kingdom. E-mail: n.j.timpson{at}bris.ac.uk.

	ABSTRACT

TOP ABSTRACT INTRODUCTION SUBJECTS AND METHODS RESULTS DISCUSSION REFERENCES

 
Background: A region of chromosome 16 containing the fat mass–and obesity-associated gene (FTO) is reproducibly associated with fat mass and body mass index (BMI), risk of obesity, and adiposity.

Objectives: We aimed to assess the possibility that appetite plays a role in the association between FTO and BMI.

Design: Detailed dietary report information from the Avon Longitudinal Study of Parents and Children allowed the exploration of relations between FTO variation and dietary intake. Analyses were performed to investigate possible associations between variation at the FTO locus and the intake of a range of micronutrients and macronutrients, with adjustment for the bias often found within dietary report data when factors related to BMI are assessed. To test the hypothesis that FTO may be influencing appetite directly, rather than indirectly via BMI and altered intake requirements, we also assessed associations between FTO and dietary intake independent of BMI.

Results: Relations between a single-nucleotide polymorphism characterizing the FTO signal (rs9939609) and dietary variables were found and can be summarized by the effect of each additional allele (per-allele effects) on total energy and total fat (P < 0.001 for both). These associations were attenuated, but they persisted specifically for fat and energy consumption after adjustment for BMI [total daily fat consumption: 1.5 g/d (P = 0.02 for the per-allele difference); total daily energy consumption: 25 kJ/d (P = 0.03 for the per-allele difference)].

Conclusion: These associations suggest that persons carrying minor variants at rs9939609 were consuming more fat and total energy than were those not carrying such variants. They also suggest that this difference was not simply dependent on having higher average BMIs among the former group.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION SUBJECTS AND METHODS RESULTS DISCUSSION REFERENCES

 
Two studies recently described an association between variation at the FTO gene locus and body mass index (BMI; in kg/m²), the risk of obesity, and adiposity (1, 2). FTO was identified as one of several loci involved in a knockout murine model resulting in a fused-toe (Ft) phenotype (3); however, it was not noted for other biological associations. Recently, work focusing on the nature of this locus showed the ubiquitous expression of this protein in fetal and adult tissues and a preferential hypothalamic mRNA profile. These findings and evidence of possible links between FTO and dietary consumption in mice (4) prompted investigation into the FTO locus as a potential influence on appetite and on its being a possible mechanism contributing to the known association between FTO and BMI or fat mass. Of particular interest is the notion that BMI-associated variation at the FTO locus may not only be associated with appetite as a result of elevated energy demand (coincident with weight increase) but may also independently influence appetite.

Recognizing fto as a demethylation catalyst, Gerken et al (4) examined fto mRNA concentrations in the hypothalamic nuclei of mice grouped by feeding behavior. This work explored whether hypothalamic fto expression is nutritionally regulated and noted that fto mRNA concentrations are indeed lower in fasted mice. This finding, along with similar patterns of hypothalamic expression and a fasting effect shown by Stratigopoulos et al (5), formed part of the first functional evidence available on the human FTO gene and provided possible links to the control of energy balance. Furthermore, the existence of variation at this locus within predicted cut-like homeobox (CUTL1) binding sites, which have demonstrable regulatory effects on the expression of FTO (and the related FTM/RPGRIP1-like locus), has suggested a route between observed genetic variation at this locus and altered hypothalamic activity (5).

Assessing the relation between BMI-associated features and appetite or dietary intake is a difficult procedure for 2 reasons. First, persons with higher BMI may have higher total energy intakes, because these are required to maintain their greater adiposity (6). Second, there may be systematic underreporting of dietary intake by persons with higher BMIs (7). A key component of the current investigation was the incorporation of a method designed to correct for this effect and to provide estimates of the effect of FTO gene variation that are not biased by these issues. In most cases, the relation between genetic variation and such anomalies is essentially randomized as a result of the Mendelian allocation of alleles at meiosis and conception (8, 9). However, in the present study, that may not be the case, because of the known association between variation at the FTO locus and BMI, in which the BMI-elevating effect of the FTO risk allele may influence energy intake or reporting tendency.

We examined the first issue through adjustment of the association between FTO and dietary intake for BMI. This approach may, of course, represent overadjustment, but the unconfounded effect should lie between the unadjusted and adjusted effect estimates. With regard to underreporting, we approached that possibility, first, by assessing the relation between genotypes and reporting accuracy and, second, by repeating the analyses, with the use of standard methods, after excluding children flagged as being underreporters (6).

The Avon Longitudinal Study of Parents and Children (ALSPAC) is a prospective study of pregnancy and childhood in which detailed dietary records have been used to assess childhood consumption of food and drink (10, 11). Considering that the FTO locus may exert an effect on adiposity through an alteration of appetite and dietary consumption, we aimed to examine the relation between genotypes at the single-nucleotide polymorphism (SNP) rs9939609 and a series of variables relating to the daily intake of dietary components. We hypothesized that the FTO allele associated with higher BMI would be related to greater energy intake, but we also carried out exploratory analyses of other dietary measures. The availability of measures of BMI and the accuracy of dietary reporting allowed us to assess whether appetite effects may be considered independent of BMI-related energy demand and also to avoid potential bias generated by BMI-related misreporting.

	SUBJECTS AND METHODS

TOP ABSTRACT INTRODUCTION SUBJECTS AND METHODS RESULTS DISCUSSION REFERENCES

 
Subjects
The ALSPAC study was conducted in a prospective cohort made up of pregnant women with expected delivery dates between April 1991 and December 1992 who were recruited from the area formerly known as the county of Avon in the United Kingdom. The cohort is population-based and broadly representative at the point of recruitment (12). More detailed information on the ALSPAC study is available at its web site (Internet: http://www.alspac.bris.ac.uk). Persons of known nonwhite racial-ethnic origin were excluded from all analyses. DNA was collected from mothers and children as described previously (13). For the present study, genotypes at SNP rs9939609 were available for 8480 singleton children. When the dataset included singleton siblings born to the same mother, only the firstborn child was included in the analyses of children.

The parents gave written informed consent to the participation of their children in the present study. All aspects of the study were reviewed and approved by the ALSPAC Law and Ethics Committee, which is registered as an Institutional Review Board. Approval was also obtained from the local research ethics committees, which are governed by the UK Department of Health.

Dietary records
Detailed dietary intake data were available for the children at age 10–11 y and were derived by using a 3-d unweighed food record. In previous work using this method of dietary assessment in a 10% subsample of this cohort, intake data were shown to compare well with the those of national surveys (10), and they had plausible relations with biological outcomes, including blood lipid concentrations (14) and concentrations of insulin-like growth factor I (11). Three-day dietary records were collected from the whole cohort between February 2002 and October 2003 when the child was aged 10–11 y. One wk before the child was due to visit a research clinic (Focus Clinic at >10 y old), a 3-d food diary (for 1 weekend day and 2 weekdays) was sent to the child to complete at home with parental help and then to bring to the clinic. A short questionnaire aimed at maintaining parental involvement and obtaining further detail regarding the foods and drinks consumed was included with the diary.

In the clinic, the diary was checked by a nutritionist, in the company of the child and usually a parent. When necessary, further detail was sought about the foods and drinks recorded, such as portion sizes and preparation methods. If the child had not completed a diary, a single 24-h recall was administered in the clinic. By this method, a response rate of 98% of clinic attendees was achieved: 13.5% had one 24-h recall–based record, and the rest had 2 d recorded. It should be noted that genotype was allocated uniformly across these persons, and there was no consistent evidence of difference in the mean values for dietary variables between these groups. The diet records were coded using Diet in, Data Out (DIDO), a coding program developed by the MRC Human Nutrition Research Unit (15) and adapted for use in coding children's diets. The coded data were converted to nutrient intakes by using a database derived from McCance and Widdowson's Composition of Foods, fifth edition (16), and supplements to that publication (17–24), augmented with manufacturers' information and information from the nutrient database used by the National Diet and Nutrition Survey (25).

Underreporting in dietary records
It is well documented that dietary data are subject to misreporting (7) and that the misreporting is usually biased toward underreporting by overweight persons (26). In a previous analysis in the subsample of ALSPAC children, we showed that the relation found between energy density of the diet at 7 y and adiposity at age 9 y was enhanced when misreporting was considered (27). This adjustment could be done either by adding misreporting status to the model or by restricting the analysis to plausible reporters. Given the principal association of genetic variation at the FTO locus with obesity or fat mass, we have therefore sought to identify participants with implausible dietary intakes by using a standard method to allow for this problem. Underreporters were defined as having a ratio of reported energy intake to predicted energy requirements of <78% (28, 29). Predicted energy requirement was calculated from body weight, after age, sex, and energy requirements for growth were taken into account (29). Underreporters were subsequently removed from analyses.

Genotyping and numbers included
Genotyping of rs9939609 was undertaken in 8480 children (1). Genotyping was performed by KBiosciences (Hoddesdon, United Kingdom) using the company's own system of fluorescence-based competitive, per-allele polymerase chain reaction (KASPar). Details of assay design are available from the KBiosciences website (Internet: http://www.kbioscience.co.uk). After adjustment for missing data and exclusion of subjects for underreporting, accurate information on dietary intake was available for 3641 children with FTO genotypes, 3589 of whom also had BMI measurements (Figure 1). Results for the overall sample, without exclusion, are reported in a supplementary table (See Table S1 under "Supplemental data" in the current online issue). The exclusion for underreporting yielded a substantial reduction in the effective sample size but was considered important because of the potentially distorting effects that misreporting subjects could have on the study's findings.

View larger version (32K): [in this window] [in a new window]   FIGURE 1. The numbers of persons included in analyses of an association between rs9939609 and dietary intake.

	RESULTS

TOP ABSTRACT INTRODUCTION SUBJECTS AND METHODS RESULTS DISCUSSION REFERENCES

 
General characteristics of the major dietary components (macronutrients) can be seen in Table 1. Energy and nutrient intakes were at levels expected for children aged 10–11 y (20); intakes of all nutrients were higher in males than in females. However, whereas consistent differences were observed between the sexes, we found no consistent differences between the dietary intakes of those with and without genotypic data (747 children) that would withstand correction for multiple testing. Thus, it appears that the children with genotypic data were representative of the whole population with regard to diet.

In an extension to this approach, despite the fact that underreporters had consistently higher BMI values than did the remaining sample [19.96 (19.81, 20.11) and 17.36 (17.29, 17.44. respectively); P = < 0.001], the present study did not show a relation between underreporting and the FTO SNP rs9939609 equivalent to that which may be predicted by the influence of FTO variation on BMI [OR = 1.01 (0.93, 1.09) for underreporter by A allele at rs9939609; P = 0.85].

Within this sample, tests of the association between rs9939609 and daily dietary consumption indicated detectable positive relations between genetic variation and the daily consumption of total energy and energy from food, saturated fat, monounsaturated fat, polyunsaturated fat (P < 0.009 for all), and trans fatty acid (P = 0.01) at age 10–11 y (Table 2 and Table 3). The effect sizes are very similar in the whole sample (see Table S1 under "Supplemental data" in the current online issue) and in the sample with exclusions for underreporting, although the significance values are greater for the whole sample, as would be expected, given the greater amount of data.

Table 4

There was no difference by genotype in the mean (and 95% CI) percentage of total energy (kcal) consumed as fat, carbohydrate, and protein [BMI-adjusted fat: 35.6% (35.5%, 35.7%)], the effect of each additional allele (per-allele effect) on percentage: 0.07 (0.09); protein: 13.6% (13.5%, 13.6%), effect on percentage: 0.06 (0.05); and carbohydrate: 50.9% (50.7%, 51.0%), per-allele effect on percentage: 0.004 (0.1)].

	DISCUSSION

TOP ABSTRACT INTRODUCTION SUBJECTS AND METHODS RESULTS DISCUSSION REFERENCES

 
We observed associations between rs9939609 and the daily intakes of energy and fat in a large representative sample of children aged 10–11 y both before and after adjustment for BMI. Frayling et al (1) reported that, by that age, the adiposity effect of this locus is stable (0.1 SDs of BMI); however, the mechanism (or mechanisms) of this relation is as yet unclear.

In the present analysis, we also were able to explore the possible effects of misreporting of dietary intake data. The removal of subjects who were flagged for underreporting appeared not to influence the distribution of genotypes at rs9939609, and, therefore, the genotype was not associated with the likelihood of a person's being either an accurate or an inaccurate reporter. Together, these findings suggest that the relatively small variance in BMI attributable to the FTO locus has not generated an association between dietary underreporting and genotype and that, when analyses are adjusted for BMI, genotype may indeed give insight into the relation between this locus and appetite.

Despite this possibility, we performed analyses after excluding persons with evidence for marked reporting bias, and we showed that SNP rs9939609 was nominally associated with the daily intakes of both fat and energy. Whereas these findings were of small effects, it was interesting that, after BMI adjustment, both our raw data and the data corrected for reporting accuracy yielded evidence of relations between FTO and dietary intake of a magnitude that is likely to be important over the life-course (30, 31). These relations include total energy, total fat, and saturated fat intakes, and they suggest that appetite may be associated with variation at this locus. However, those results are likely to reflect portions of the variance in dietary intake that are explained by both basal energy demands (which vary by BMI) and a possible direct effect of the FTO locus on food intake.

An important consideration in the interpretation of these results is that of the possibility of overadjustment as a result of the incorporation of BMI into models of the association between variation at rs9939609 and appetite. BMI was incorporated into the analysis of relations between variation at the FTO locus and dietary intake in an effort to remove the effects of the known correlation between metabolic requirements and BMI in accurate dietary reporters. However, because BMI is in part an outcome of dietary intake, such an adjustment may attenuate appetite effects. We consider it likely that the actual point estimates of the effect of FTO on dietary intake will lie between the estimates presented before this adjustment and the estimates presented afterward.

It is important that the association between rs9939609 variation and the intake of energy or fat also may be attenuated by the known imprecision of dietary intake measurement and the limited extent to which short-term dietary measures reflect long-term patterns. Whereas this attenuation is a limitation, the observations in the present study still add to the weight of evidence for a BMI-independent relation between FTO and appetite.

Analysis by effect on the percentage of total energy derived from specific food types indicated that, if the observed appetite effect is real, then it is not restricted to any particular dietary component. This possibility would, therefore, suggest that the observed relation is one of a generic effect on food intake.

The present analysis makes use of the best available evidence to date for the assessment of a possible direct role of the FTO locus in regulating dietary intake. In the case of total energy and fat, we have observed nominal evidence for a BMI-independent association between this locus and intake, even after considerable reduction in sample size as a result of the removal of apparently underreporting subjects. This removal did not change the magnitude of the association, but it necessarily reduced statistical precision. The independent appetite effect is small, but it is of a magnitude that, over the life-course, would be likely to lead to differences in BMI at least approaching those due to FTO genotype (32–34).

Greater numbers of participants or more accurate assessments of dietary intake (or both) are required for a more comprehensive assessment of this relation. However, our data are consistent with data from other sources (35) in suggesting that small changes in energy intake will, as they accumulate over time, lead to substantial changes in BMI.

	ACKNOWLEDGMENTS

 
We are extremely grateful to all the families who took part in this study, the midwives for their help in recruiting them, and the whole ALSPAC team, which includes interviewers, computer and laboratory technicians, clerical workers, research scientists, volunteers, managers, receptionists, and nurses.

The authors' responsibilities were as follows—NJT: coordinated the present study and undertook the analysis and writing of the manuscript; PE: assisted in the writing of the manuscript; IR: assisted in the organization of data and in the writing of the manuscript; TMF, ATH, and MIM: integral parts of the initial analysis team for FTO and helped develop the hypotheses for the study; and GDS (the Principal Investigator responsible for the ALSPAC cohort): helped to generate the hypothesis for the study and assisted in the writing of the manuscript. This publication is the work of the authors, who serve as guarantors for the contents of this paper. None of the authors had a personal or financial conflict of interest.

	REFERENCES

TOP ABSTRACT INTRODUCTION SUBJECTS AND METHODS RESULTS DISCUSSION REFERENCES

 

1. Frayling TM, Timpson NJ, Weedon MN, et al. A common variant in the FTO gene is associated with body mass index and predisposes to childhood and adult obesity. Science 2007;316:889–94.[Abstract/Free Full Text]
2. Dina C, Meyre D, Gallina S, et al. Variation in FTO contributes to childhood obesity and severe adult obesity. Nat Genet 2007;39:724–6.[Medline]
3. Anselme I, Laclef C, Lanaud M, Ryther U, Schneider-Maunoury S. Defects in brain patterning and head morphogenesis in the mouse mutant fused toes. Develop Biol 2007;304:208–20.[Medline]
4. Gerken T, Girard CA, Tung YL, et al. The obesity-associated FTO gene encodes a 2-oxoglutarate-dependent nucleic acid demethylase. Science 2007;318:1469–71 (Epub 2007 Nov 8).[Abstract/Free Full Text]
5. Stratigopoulos G, Padilla SL, LeDuc CA, et al. Regulation of Fto/Ftm gene expression in mice and humans. Am J Physiol Regul Integr Comp Physiol 2008;294:R1185–96 (Epub 2008 Feb 6).[Abstract/Free Full Text]
6. Rennie K, Coward WA, Jebb SA. Estimating under-reporting of energy intake in dietary surveys using and individualised method. Br J Nutr 2007;97:1169–76.[Medline]
7. Bingham S, Welch A, Cassidy A, Runswick S. The use of 24 h urine nitrogen to detect bias in the reported habitual food intake of individuals assessed from weighed dietary records. Proc Nutr Soc 1991;50:32A (abstr).
8. Davey Smith G, Ebrahim S. ‘Mendelian randomisation’: can genetic epidemiology contribute to understanding environmental determinants of disease? Int J Epidemiol 2003;32:1–22.[Abstract/Free Full Text]
9. Davey Smith G, Ebrahim S. Mendelian randomisation: prospects, pitfalls and limitations. Int J Epidemiol 2004;33:30–42.[Free Full Text]
10. Emmett P, Rogers I, Symes C; ALSPAC Study Team. Avon Longitudinal Study of Pregnancy and Childhood. Food and nutrient intakes of a population sample of 3-year-old children in the South West of England in 1996. Public Health Nutr 2002;5:55–64.[Medline]
11. Rogers I, Gunnell D, Emmett P, et al. Cross-sectional associations of diet and insulin-like growth factor levels in 7- to 8-year-old children. Cancer Epidemiol Biomarkers Prev 2005;14:204–12.[Abstract/Free Full Text]
12. Golding J, Pembrey M, Jones R; ALSPAC Study Team. ALSPAC—the Avon Longitudinal Study of Parents and Children. I. Study methodology. Paediatr Perinat Epidemiol 2001;15:74–87.[Medline]
13. Jones RW, Ring S, Tyfield L, et al. A new human genetic resource: a DNA bank established as part of the Avon longitudinal study of pregnancy and childhood (ALSPAC). Eur J Hum Genet 2000;8:653–60.[Medline]
14. Cowin I, Emmett P; ALSPAC Study Team. Association between dietary intakes and blood cholesterol concentrations at 31 months. Eur J Clin Nutr 2001;55:39–49.[Medline]
15. Price GM, Paul AA, Key FB, et al. Measurement of diet in a large national survey: comparison of computerised and manual coding of records in household measures. J Hum Nutr Diet 1995;8:417–28.
16. Holland B, Welch AA, Unwin ID, Buss DH, Paul AA, Southgate DAT. McCance and Widdowson's The composition of foods, 5th ed. Cambridge, United Kingdom: Royal Society of Chemistry, 1991.
17. Holland B, Unwin ID, Buss DH. Cereals and cereal products. Third supplement to McCance and Widdowson's The composition of foods, 4th ed. Cambridge, United Kingdom: Royal Society of Chemistry, 1998.
18. Holland B, Brown J, Buss DH. Fish and fish products. Third supplement to McCance and Widdowson's The composition of foods, 5th ed. Cambridge, United Kingdom: Royal Society of Chemistry, 1993.
19. Holland B, Unwin ID, Buss DH. Vegetables, herbs and spices. Fifth supplement to McCance and Widdowson's The composition of foods, 4th ed. Cambridge, United Kingdom: Royal Society of Chemistry, 1991.
20. Holland B, Unwin ID, Buss DH. Fruit and nuts. First supplement to McCance and Widdowson's The composition of foods, 5th ed. Cambridge, United Kingdom: Royal Society of Chemistry, 1992.
21. Holland B, Unwin ID, Buss DH. Milk products and eggs. Fourth supplement to McCance and Widdowson's The composition of foods, 4th ed. Cambridge, United Kingdom: Royal Society of Chemistry, 1989.
22. Chan W, Brown J, Church SM, Buss DH. Meat products and dishes. Sixth supplement to McCance and Widdowson's The composition of foods, 5th ed. Cambridge, United Kingdom: Royal Society of Chemistry, 1996.
23. Chan W, Brown J, Buss DH. Miscellaneous foods. Fourth supplement to McCance and Widdowson's The composition of foods, 5th ed. Cambridge, United Kingdom: Royal Society of Chemistry, 1994.
24. Chan W, Brown J, Lee SM, Buss DH. Meat, poultry and game. Fifth supplement to McCance and Widdowson's The composition of foods, 5th ed. Cambridge, United Kingdom: Royal Society of Chemistry, 1995.
25. Gregory J, Lowe S. National Diet and Nutrition Survey. Young people aged 4 to 18 years. London: The Stationery Office 2000.
26. Livingstone MBE, Robson PJ, Wallace JMW. Issues in dietary intake assessment of children and adolescents. Br J Nutr 2004;92:S213–22.[Medline]
27. Johnson L, Mander AP, Jones LR, Emmett P, Jebb SA. A prospective analysis of dietary energy density at age 5 and 7 years and fatness at 9 years among UK children. Int J Obes (Lond) 2008;32:586–93 (Epub 2007 Oct 2).[Medline]
28. Black AE, Cole TJ. Within and between subject variation in energy expenditure measured by the doubly-labelled water technique: implications for validated reported dietary energy intake. Eur J Clin Nutr 2000;54:386–94.[Medline]
29. Torun B. Energy requirements of children and adolescents. Public Health Nutr 2005;8:968–93.[Medline]
30. Bray GA. Risks of obesity. Primary Care 2003;30:281–99.[Medline]
31. Bray GA, Popkin BM. Dietary fat intake does affect obesity! Am J Clin Nutr 1998;68:1157–73.[Abstract]
32. Berkey CS, Rockett HR, Field AE, et al. Activity, dietary intake, and weight changes in a longitudinal study of preadolescent and adolescent boys and girls. Pediatrics 2000;105:E56.[Medline]
33. Klesges RC, Klesges LM, Eck LH, Shelton ML. A longitudinal analysis of accelerated weight gain in preschool children. Pediatrics 1995;95:126–30.[Abstract/Free Full Text]
34. Rosenbaum M, Leibel RL. The physiology of body weight regulation: relevance to the etiology of obesity in children. Pediatrics 1998;101:525–39.[Abstract/Free Full Text]
35. Hill JO, Wyatt HR, Reed GW, Peters JC. Obesity and the environment: where do we go from here? Science 2003;299:853–5.[Abstract/Free Full Text]

This Article Abstract Full Text (PDF) Supplemental data Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Google Scholar Articles by Timpson, N. J Articles by Davey Smith, G. PubMed PubMed Citation Articles by Timpson, N. J Articles by Davey Smith, G. Agricola Articles by Timpson, N. J Articles by Davey Smith, G.