OBJECTIVE

FTO is the most important polygene identified for obesity. We aimed to investigate whether a variant in FTO affects type 2 diabetes risk entirely through its effect on BMI and how FTO influences BMI across adult life span.

RESEARCH DESIGN AND METHODS

Through regression models, we assessed the relationship between the FTO single nucleotide polymorphisms rs9939609, type 2 diabetes, and BMI across life span in subjects from the Norwegian population-based HUNT study using cross-sectional and longitudinal perspectives. For replication and meta-analysis, we used data from the Malmö Diet and Cancer (MDC) and Malmö Preventive Project (MPP) cohorts, comprising a total sample of 41,504 Scandinavians.

RESULTS

The meta-analysis revealed a highly significant association for rs9939609 with both type 2 diabetes (OR 1.13; P = 4.5 × 10−8) and the risk to develop incident type 2 diabetes (OR 1.16; P = 3.2 × 10−8). The associations remained also after correction for BMI and other anthropometric measures. Furthermore, we confirmed the strong effect on BMI (0.28 kg/m2 per risk allele; P = 2.0 × 10−26), with no heterogeneity between different age-groups. We found no differences in change of BMI over time according to rs9939609 risk alleles, neither overall (∆BMI = 0.0 [−0.05, 0.05]) nor in any individual age stratum, indicating no further weight gain attributable to FTO genotype in adults.

CONCLUSIONS

We have identified that a variant in FTO alters type 2 diabetes risk partly independent of its observed effect on BMI. The additional weight gain as a result of the FTO risk variant seems to occur before adulthood, and the BMI difference remains stable thereafter.

Genomewide association studies (GWAS) have identified a strong correlation between BMI and FTO single nucleotide polymorphisms (SNPs) (14), and the association has been confirmed in multiple populations (reviewed in 5). The FTO risk variants are also associated with obesity-related traits (68). However, these effects appear to be secondary to weight increase because the associations are attenuated after adjusting for BMI (2). In contrast, we and others have found that the association with type 2 diabetes may not be completely mediated through BMI, because it remains significant after BMI correction (9). This indicates that the relationship between sequence variation in FTO and type 2 diabetes is not fully mediated through BMI or that BMI in some populations does not reveal accurate estimates of the effect of FTO on adiposity.

Various studies have investigated the effect of FTO variants on BMI and weight in a longitudinal perspective (1018) but with diverging results. With access to extensive data from three large Scandinavian populations, through a meta-analysis approach using both cross-sectional and longitudinal data, we aimed to investigate whether the FTO risk allele affects type 2 diabetes risk after correction for BMI and whether it influences weight gain during adult life.

Definition of cohorts.

We studied HUNT2, a subset (aged ≥20) of a Norwegian population-based health survey (Nord-Trøndelag Health Study) (19). Our material comprised 1,740 diabetic individuals (1,543 with type 2 diabetes) and 3,856 population-based control subjects drawn from the same study population. We also had access to data on diabetes status, weight, and height from HUNT1 (1985) for 4,625 of the 5,596 subjects in HUNT2 (1995), i.e., 10-year follow-up. During these 10 years, 1,089 individuals developed type 2 diabetes. Diagnosis of diabetes was self-reported or identified by standard tests if random glucose was >8.0 mmol/L.

The Malmö Diet and Cancer (MDC) cohort (20) with baseline examinations from 1991 to 1996 consisted of 28,449 individuals. All men born between 1923 and 1945 and all women born between 1923 and 1950 from Malmö were invited. Diabetes diagnosis at baseline was self-reported or diagnosed if fasting plasma glucose was ≥7.0 mmol/L.

In the Malmö Preventive Project (MPP) cohort (21), 33,346 subjects from Malmö participated in a health screening. Men were included from 1974 to 1990, and women were included from 1980 to 1992. Eligible participants (25,000) were invited to a rescreening visit during 2002–2006. Of those invited, 16,061 nondiabetic subjects, 2,063 of whom developed type 2 diabetes during follow-up, were included in the current study. Diabetes diagnosis was taken from patient records or if fasting plasma glucose was ≥7.0 mmol/L.

The clinical characteristics of individuals from the three cohorts are shown in Table 1.

TABLE 1

Clinical characteristics of the individuals from the three different cohorts

HUNT
MPP
MDC
NAllNType 2 diabetesNNo type 2 diabetesNAllNType 2 diabetesNType 2 diabetesNAllNType 2 diabetesNType 2 diabetes
Individuals (n5,596 — 1,543 — 4,053 — 15,930 — 2,054 — 13,876 — 19,978 — 720 — 19,258 — 
Sex (male/female) 5,596 2,788/2,808 1,543 726/817 4,053 2,062/1,991 15,930 10,335/5,595 2,054 1,604/450 13,876 8,731/5,145 19,978 6,529/13,449 720 364/356 19,258 6,165/13,093 
Age (years) 5,596 59.8 ± 17.1 1,543 67.9 ± 12.0 4,053 56.7 ± 17.8 15,930 45.5 ± 6.9 2,054 45.0 ± 6.3 13,876 45.6 ± 7.0 19,978 57.7 ± 8.4 720 61.3 ± 7.3 19,258 57.6 ± 8.5 
Follow-up time (years) 4,625 10 1421 10 3,204 10 15,930 23.5 ± 4.2 2,054 24.5 ± 3.4 13,876 23.4 ± 4.3 — — — — — — 
BMI baseline (kg/m24,625 26.8 ± 4.4 1,421 29.5 ± 4.8 3,204 25.6 ± 3.7 15,925 24.3 ± 3.3 2,054 26.2 ± 3.8 13,871 24.0 ± 3.1 19,978 25.8 ± 4.1 720 28.1 ± 4.7 19,258 25.7 ± 4.1 
BMI (kg/m25,596 27.3 ± 4.6 1543 29.3 ± 4.8 4,053 26.6 ± 4.1 15,843 27.1 ± 4.1 2,038 29.3 ± 4.6 13,805 26.8 ± 3.9 — — — — — — 
Waist-to-hip ratio 5,552 0.86 ± 0.1 1,530 0.89 ± 0.1 4,022 0.86 ± 0.1 15,820 0.92 ± 0.1 2,030 0.97 ± 0.1 13,790 0.91 ± 0.1 19,959 0.85 ± 0.1 717 0.9 ± 0.1 19,242 0.84 ± 0.1 
Waist circumference (cm) 5,552 90.3 ± 12.2 1,530 95.8 ± 11.8 4,022 88.2 ± 11.6 15,829 94.8 ± 12.2 2,030 102.1 ± 12.1 13,799 93.7 ± 11.8 19,963 83.5 ± 13.2 718 93.4 ± 13.8 19,245 83.2 ± 13.0 
Serum triglycerides (mmol/L) 5,583 2.0 ± 1.3 1,534 2.5 ± 1.6 4,049 1.8 ± 1.1 15,921 1.3 ± 0.8 2,052 1.6 ± 1.2 13,869 1.2 ± 0.7 3,228 1.4 ± 0.9 124 2.2 ± 1.5 3,104 1.4 ± 0.8 
Serum cholesterol (mmol/L) 5,584 6.1 ± 1.3 1,534 6.2 ± 1.3 4,050 6.1 ± 1.3 15,923 5.6 ± 1.1 2,052 5.2 ± 1.1 13,871 5.7 ± 1.1 3,229 6.2 ± 1.1 124 6.2 ± 1.0 3,105 6.2 ± 1.1 
Serum HDL (mmol/L) 5,582 1.3 ± 0.4 1,533 1.2 ± 0.4 4,049 1.4 ± 0.4 15,915 1.4 ± 0.4 2,052 1.2 ± 0.4 13,863 1.4 ± 0.4 3,188 1.4 ± 0.4 119 1.2 ± 0.4 3,069 1.4 ± 0.4 
Fasting plasma glucose (mmol/L)* 5,582 6.8 ± 3.3 1,533 9.6 ± 4.2 4,049 5.7 ± 1.9 15,921 5.8 ± 1.3 2,050 7.9 ± 2.4 13,871 5.5 ± 0.5 3,228 6.0 ± 1.9 125 12.3 ± 4.7 3,103 5.7 ± 1.0 
HUNT
MPP
MDC
NAllNType 2 diabetesNNo type 2 diabetesNAllNType 2 diabetesNType 2 diabetesNAllNType 2 diabetesNType 2 diabetes
Individuals (n5,596 — 1,543 — 4,053 — 15,930 — 2,054 — 13,876 — 19,978 — 720 — 19,258 — 
Sex (male/female) 5,596 2,788/2,808 1,543 726/817 4,053 2,062/1,991 15,930 10,335/5,595 2,054 1,604/450 13,876 8,731/5,145 19,978 6,529/13,449 720 364/356 19,258 6,165/13,093 
Age (years) 5,596 59.8 ± 17.1 1,543 67.9 ± 12.0 4,053 56.7 ± 17.8 15,930 45.5 ± 6.9 2,054 45.0 ± 6.3 13,876 45.6 ± 7.0 19,978 57.7 ± 8.4 720 61.3 ± 7.3 19,258 57.6 ± 8.5 
Follow-up time (years) 4,625 10 1421 10 3,204 10 15,930 23.5 ± 4.2 2,054 24.5 ± 3.4 13,876 23.4 ± 4.3 — — — — — — 
BMI baseline (kg/m24,625 26.8 ± 4.4 1,421 29.5 ± 4.8 3,204 25.6 ± 3.7 15,925 24.3 ± 3.3 2,054 26.2 ± 3.8 13,871 24.0 ± 3.1 19,978 25.8 ± 4.1 720 28.1 ± 4.7 19,258 25.7 ± 4.1 
BMI (kg/m25,596 27.3 ± 4.6 1543 29.3 ± 4.8 4,053 26.6 ± 4.1 15,843 27.1 ± 4.1 2,038 29.3 ± 4.6 13,805 26.8 ± 3.9 — — — — — — 
Waist-to-hip ratio 5,552 0.86 ± 0.1 1,530 0.89 ± 0.1 4,022 0.86 ± 0.1 15,820 0.92 ± 0.1 2,030 0.97 ± 0.1 13,790 0.91 ± 0.1 19,959 0.85 ± 0.1 717 0.9 ± 0.1 19,242 0.84 ± 0.1 
Waist circumference (cm) 5,552 90.3 ± 12.2 1,530 95.8 ± 11.8 4,022 88.2 ± 11.6 15,829 94.8 ± 12.2 2,030 102.1 ± 12.1 13,799 93.7 ± 11.8 19,963 83.5 ± 13.2 718 93.4 ± 13.8 19,245 83.2 ± 13.0 
Serum triglycerides (mmol/L) 5,583 2.0 ± 1.3 1,534 2.5 ± 1.6 4,049 1.8 ± 1.1 15,921 1.3 ± 0.8 2,052 1.6 ± 1.2 13,869 1.2 ± 0.7 3,228 1.4 ± 0.9 124 2.2 ± 1.5 3,104 1.4 ± 0.8 
Serum cholesterol (mmol/L) 5,584 6.1 ± 1.3 1,534 6.2 ± 1.3 4,050 6.1 ± 1.3 15,923 5.6 ± 1.1 2,052 5.2 ± 1.1 13,871 5.7 ± 1.1 3,229 6.2 ± 1.1 124 6.2 ± 1.0 3,105 6.2 ± 1.1 
Serum HDL (mmol/L) 5,582 1.3 ± 0.4 1,533 1.2 ± 0.4 4,049 1.4 ± 0.4 15,915 1.4 ± 0.4 2,052 1.2 ± 0.4 13,863 1.4 ± 0.4 3,188 1.4 ± 0.4 119 1.2 ± 0.4 3,069 1.4 ± 0.4 
Fasting plasma glucose (mmol/L)* 5,582 6.8 ± 3.3 1,533 9.6 ± 4.2 4,049 5.7 ± 1.9 15,921 5.8 ± 1.3 2,050 7.9 ± 2.4 13,871 5.5 ± 0.5 3,228 6.0 ± 1.9 125 12.3 ± 4.7 3,103 5.7 ± 1.0 

Data are presented as means ± SD. Data presented for the HUNT and MPP cohorts are follow-up measures unless otherwise stated. All data presented for the MDC cohort are baseline measures as a result of no available follow-up measures.

*Only nonfasting glucose measures were available for participants in the HUNT cohort.

SNP selection, genotyping, and quality control.

Because the reported obesity-associated FTO SNPs are in strong to perfect linkage disequilibrium (LD) (pairwise r2 > 0.8; HapMap; CEU, release 21), we included rs9939609, reported by the first GWAS (2), as proxy. We genotyped HUNT2 subjects by MassARRAY iPLEX System (SEQUENOM, San Diego, CA). Duplicate concordance rate was 99.7% (N = 3,761). MDC individuals were genotyped using MassARRAY iPLEX System or a TaqMan assay (Applied Biosystems, Foster City, CA). Individuals (249 total) were genotyped with both methods (99.4% allelic concordance rate). A total of 8,175 individuals were genotyped twice (allelic concordance rate 99.3%). MPP subjects were genotyped by TaqMan assay with duplicate concordance rate 99.7% (N = 7,926). The final genotyping success rates were 99.7, 99.2, and 97.2%, in HUNT2, MPP, and MDC, respectively. The FTO SNP was in Hardy-Weinberg equilibrium (P value >0.05) in all cohorts.

Statistical analysis.

We used logistic regression to investigate the association between type 2 diabetes and FTO genotype under an additive model. Age, sex, different combinations of BMI, waist circumference, and waist-to-hip ratio were included as covariates. In HUNT2 and MPP, follow-up measures were used unless otherwise stated. In MDC, baseline measures were used because of no available follow-up. To evaluate the risk of developing incident type 2 diabetes according to the three FTO genotypic classes, we designed a case-control study using only subjects who were healthy at baseline. Those developing type 2 diabetes during follow-up were defined as cases and the rest as control subjects. We used logistic regression models corrected for sex, baseline age, and BMI. In a second step, we included the change in BMI over time (∆BMI) as an additional cofactor.

To analyze the association between FTO genotype and BMI, we used linear regression models assuming additive effects of allele dosage with adjustment for age, sex, and diabetes status. Logarithmic transformation of BMI values did not change the results; thus, only results using nontransformed BMI are presented. To assess whether the allele-wise increase in BMI differed across different adult ages, we performed individual analysis in every 10-year age stratum in the cross-sectional and longitudinal datasets. Finally, meta-analysis was performed to combine the regression coefficients (per allele change in BMI) with their standard errors from within the three cohorts and for each specific age-group. Interstudy heterogeneity and heterogeneity between different age-groups were estimated using Cochran’s Q test and the I2 statistic. Overall estimates were calculated using a fixed-effect model with inverse variance.

We performed statistical analyses by PLINK (22) and Stata SE v10.0 (Stata, Brownsville, TX) or SPSS (version 18; SPSS Inc., Chicago, IL). Meta-analysis statistics and plots were produced using the METAN module (23) developed for Stata and with GWAMA (Genome-Wide Association Meta-Analysis) software (24).

Relationship among FTO, type 2 diabetes, and obesity-related quantitative traits across life span in HUNT.

After correction for age and sex, we observed a strong association with type 2 diabetes for rs9939609 in HUNT2. This association remained significant after correction for BMI (OR 1.19 [95%CI 1.09–1.30]; P = 1.8 × 10−5). The FTO variant also conferred an increased risk for type 2 diabetes after adjustment for waist circumference and waist-to-hip ratio. These results suggested that rs9939609 has an effect on the risk of type 2 diabetes, an effect that cannot be entirely explained through increased BMI or central obesity.

Using a cross-sectional design, we observed that the FTO-associated allele-wise increase in BMI persisted at the same level throughout life (Supplementary Fig. 1). In addition, rs9939609 × age interactions on obesity-related traits were all nonsignificant (Supplementary Table 1), suggesting that changes in these traits by age were not dependent on the individual’s FTO genotype.

Next, we studied longitudinal change in the association between FTO and BMI during 10-year follow-up from HUNT1 to HUNT2. The FTO variant showed an association with all obesity-related quantitative traits (Supplementary Table 1). There was, however, no association between rs9939609 and ∆BMI between 1985 and 1995. This suggested that the FTO-associated relative difference of BMI is established before adulthood and then remains stable.

Confirmation of the findings from HUNT–meta-analysis in 41,504 Scandinavians.

Clinical characteristics of individuals from the three different cohorts are presented in Table 1. The minor allele frequencies of rs9939609 in nondiabetic individuals were 0.42, 0.41, and 0.41 in HUNT2, MPP, and MDC, respectively.

The meta-analysis demonstrated that the association between rs9939609 and type 2 diabetes was strong after adjustment for age and sex (OR 1.13 [95%CI 1.08–1.19]; P = 4.5 × 10−8) and remained significant after BMI correction (OR 1.09 [95%CI 1.04–1.15]; P = 1.2 × 10−4; Fig. 1A and B). Correction for waist-to-hip ratio or waist circumference instead of BMI did not change the results (Supplementary Fig. 2A–C). To further elucidate whether rs9939609 exerts an effect on type 2 diabetes independently of BMI, we evaluated the risk to develop incident type 2 diabetes according to FTO genotype during follow-up. As shown in Supplementary Fig. 3A–C, the association remained similar for incident type 2 diabetes after correction for sex and baseline age and BMI (OR 1.12 [95%CI 1.05–1.18]; P = 1.1 × 10−4) and after correction also for ∆BMI (OR 1.11 [95%CI 1.05–1.18]; P = 1.5 × 10−4).

FIG. 1.

Meta-analysis plots of association between FTO and type 2 diabetes comprising 4,317 subjects with type 2 diabetes and 37,187 control subjects. A: Meta-analysis plot of association between FTO rs9939609 and type 2 diabetes after correction for age and sex (allelic OR 1.13 [95%CI 1.08–1.19]). B: Meta-analysis plot of association between FTO rs9939609 and type 2 diabetes after correction for age, sex, and BMI (allelic OR 1.09 [ 1.04–1.15]). We observed a tendency toward heterogeneity between the samples (P = 0.064), and the variation in the estimate attributable to heterogeneity was calculated to 63.7%. The ORs for the overall estimates were calculated using a fixed-effect model with inverse variance. The weighting (%weight) represents the inverse variance of each studies’ effect estimator.

FIG. 1.

Meta-analysis plots of association between FTO and type 2 diabetes comprising 4,317 subjects with type 2 diabetes and 37,187 control subjects. A: Meta-analysis plot of association between FTO rs9939609 and type 2 diabetes after correction for age and sex (allelic OR 1.13 [95%CI 1.08–1.19]). B: Meta-analysis plot of association between FTO rs9939609 and type 2 diabetes after correction for age, sex, and BMI (allelic OR 1.09 [ 1.04–1.15]). We observed a tendency toward heterogeneity between the samples (P = 0.064), and the variation in the estimate attributable to heterogeneity was calculated to 63.7%. The ORs for the overall estimates were calculated using a fixed-effect model with inverse variance. The weighting (%weight) represents the inverse variance of each studies’ effect estimator.

Close modal

The meta-analysis of the FTO-associated allele-wise effect on BMI using cross-sectional data confirmed the strong effect of the FTO SNP on BMI (0.28 kg/m2 per risk allele [P = 2.0 × 10−26]; Fig. 2A). Furthermore, we detected no heterogeneity in the effect sizes for the FTO risk allele between the different age-groups (Fig. 2B). Finally, Fig. 3 shows the linear regression summary results between rs9939609 and ∆BMI for all HUNT and MPP individuals for whom longitudinal data were available. There was no significant difference in ∆BMI according to overall number of rs9939609 risk alleles (∆BMI = 0.0 [−0.05, 0.05]) or in any individual age stratum (Fig. 3B). Hence, the FTO-associated effect on BMI seems to establish relatively early in life, and the relative BMI difference remains stable across adult life.

FIG. 2.

Meta-analysis plot of the FTO-associated allele-wise effect on BMI using cross-sectional data. The results included in the meta-analysis are from regression analysis adjusted for age, sex, and diabetes status. The weighting (%weight) represents the inverse variance of each study’s effect estimator. A: Meta-analysis plot comprising all 41,504 individuals. No heterogeneity between the cohorts was detected (P = 0.242), and the overall allelic effect was estimated to 0.28 kg/m2. B: Meta-analysis plot comprising all 41,504 individuals stratified on 10-year age strata. No heterogeneity between the subgroups was detected (P = 0.378). Moderate heterogeneity was, however, observed in two of the subgroups.

FIG. 2.

Meta-analysis plot of the FTO-associated allele-wise effect on BMI using cross-sectional data. The results included in the meta-analysis are from regression analysis adjusted for age, sex, and diabetes status. The weighting (%weight) represents the inverse variance of each study’s effect estimator. A: Meta-analysis plot comprising all 41,504 individuals. No heterogeneity between the cohorts was detected (P = 0.242), and the overall allelic effect was estimated to 0.28 kg/m2. B: Meta-analysis plot comprising all 41,504 individuals stratified on 10-year age strata. No heterogeneity between the subgroups was detected (P = 0.378). Moderate heterogeneity was, however, observed in two of the subgroups.

Close modal
FIG. 3.

Meta-analysis plot of the FTO-associated effect on BMI differences using longitudinal data from the HUNT and MPP study. The results included in the meta-analysis are from regression analysis adjusted for age, sex, and diabetes status. The weighting (%weight) represents the inverse variance of each study’s effect estimator. A: Meta-analysis plot comprising all 20,464 individuals with follow-up data on BMI. No heterogeneity between the cohorts was detected (P = 0.892), and the overall allelic effect for the FTO SNP on BMI difference over a period of time was estimated to 0 kg/m2. B: Meta-analysis plot comprising all 20,464 individuals stratified on 10-year age strata. Each age stratum reflects the age at baseline. No heterogeneity between the subgroups was detected (P = 0.967). Moderate heterogeneity was, however, observed in two of the subgroups.

FIG. 3.

Meta-analysis plot of the FTO-associated effect on BMI differences using longitudinal data from the HUNT and MPP study. The results included in the meta-analysis are from regression analysis adjusted for age, sex, and diabetes status. The weighting (%weight) represents the inverse variance of each study’s effect estimator. A: Meta-analysis plot comprising all 20,464 individuals with follow-up data on BMI. No heterogeneity between the cohorts was detected (P = 0.892), and the overall allelic effect for the FTO SNP on BMI difference over a period of time was estimated to 0 kg/m2. B: Meta-analysis plot comprising all 20,464 individuals stratified on 10-year age strata. Each age stratum reflects the age at baseline. No heterogeneity between the subgroups was detected (P = 0.967). Moderate heterogeneity was, however, observed in two of the subgroups.

Close modal

To our knowledge, this is the largest study investigating the effect of FTO sequence variants on type 2 diabetes and BMI across the whole range of adult ages and in a longitudinal perspective. In 41,504 Scandinavians, we demonstrate that a common variant of FTO does not mediate type 2 diabetes risk entirely through its influence on BMI. Although our findings are comparable with some earlier studies (2527), they contrast previous results reported in most populations studied to date, including Europeans (13,8). Reasons for the diverging results could be differences in selection or recruitment of cases and control subjects between studies, differences in undetected key effects at early age, or population-specific environmental factors that may interact with the way FTO works to influence the risk of type 2 diabetes. In an attempt to capture the complex relationship between FTO, BMI, and type 2 diabetes during the life course, we performed an analysis on incident type 2 diabetes. The results remained similar in the longitudinal study both when we controlled for BMI at baseline (before diabetes was diagnosed), ∆BMI, or waist circumference and/or waist-to-hip ratio as covariates in the regression analyses. None of the covariates alone or in combination with BMI changed our results notably. FTO still conferred an increased risk for type 2 diabetes.

How sequence variation in FTO could possibly affect type 2 diabetes risk in other forms than through increased adiposity remains elusive. No associations have been reported between FTO SNPs and glucose tolerance or insulin sensitivity. A link between SNPs in FTO and altered lipid profiles has been suggested (6,9), but we could not confirm this in our meta-analysis (Supplementary Table 2). It has been suggested that rs9939609 affects the primary allelic FTO transcript levels (28), and correlations have been observed in peripheral tissues between BMI of tissue donors and FTO mRNA expression levels (29). It is noteworthy that three recent FTO expression studies support a potential role in type 2 diabetes independently of BMI. One study found no association between FTO expression and BMI in islet cells (30). Another study reported an inverse correlation between Fto mRNA and glucose in mice after correction for body weight (31). Finally, a third study found an increase of FTO mRNA and protein levels in muscle from type 2 diabetic patients compared with healthy lean control subjects or BMI-matched obese nondiabetic individuals (32). The latter also suggests that increased FTO expression in type 2 diabetic patients contributes to reduced mitochondria oxidative capacities, lipid accumulation, and oxidative stress, all associated with type 2 diabetes. It is also possible that the rs9939609 SNP (or a SNP in strong LD) affects another gene in the region, which has the potential to alter type 2 diabetes risk independently of BMI (33).

The association between FTO sequence variants and BMI is not established at birth (2,34) but seems to evolve gradually before adulthood (2,35,36). It is not clear how FTO genotype affects BMI after adolescence and develops during the life course (1018), although a recent longitudinal Finnish study suggests that the effect may continue into adulthood since they found an association between rs9939609 and BMI at age 31, which could not be explained by the BMI at age 14 (18). Using cross-sectional and longitudinal designs, we identified in the three Scandinavian populations that the relative difference in mean BMI among individuals with different rs9939609 genotypes remains surprisingly stable across all adult ages. Hence, because our study primarily comprised individuals that were above 30 years of age (98.7%), current evidence suggest that the FTO variant increases BMI in the first 2 to 3 decades of life, and from then on the BMI difference between the genotypes becomes more or less constant throughout life. Nevertheless, it remains to be seen whether other relevant factors such as diet and physical activity may interact and modify the susceptibility to obesity by the FTO variants during the life course (3739).

In summary, we have replicated that a common variant in the FTO gene alters type 2 diabetes risk but find that this association is partly independent of the effect on BMI. Our data further demonstrate that the weight gain as a result of the FTO risk variant occurs during youth and that the BMI difference according to the FTO genotype persists at the same level throughout life, setting the threshold for BMI.

J.K.H. and S.J. contributed equally to this work.

E.S. and A.J. contributed equally to this work.

The HUNT study was supported in part by funds from the University of Bergen, Haukeland University Hospital, Helse Vest, Innovest, and the Research Council of Norway. Genotyping was in part provided by the CIGENE technology platform (Ås, Norway), which is supported by the Functional Genomics Programme (FUGE) of the Research Council of Norway. HUNT is a collaboration between the HUNT Research Center at the Norwegian University of Science and Technology, Levanger; the Norwegian Institute for Public Health; and the Nord-Trøndelag County Council. The diabetes part of HUNT was partly supported by funds from GlaxoSmithKline Norway and the Norwegian Diabetes Association. The MDC study was supported by project grants from the Swedish Research Council, the European Foundation for the Study of Diabetes, the Novo Nordisk and Albert Påhlsson Foundations, a Linnaeus grant to the Lund University Diabetes Centre, and the Knut and Alice Wallenberg Foundation. The MPP study was supported by grants from the Swedish Research Council (including Linné grant 31475113580), the Heart and Lung Foundation, the Diabetes Research Society, a Nordic Center of Excellence Grant in Disease Genetics, the Diabetes Program at the Lund University, the European Foundation for the Study of Diabetes, the Påhlsson Foundation, the Craaford Foundation, the Novo Nordisk Foundation, the European Network of Genomic and Genetic Epidemiology, and the Wallenberg Foundation. L.G. has been a consultant for and served on advisory boards for sanofi-aventis, GlaxoSmithKline, Novartis, Merck, Tethys Bioscience, and Xoma and received lecture fees from Eli Lilly and Novartis. No other potential conflicts of interest relevant to this article were reported.

J.K.H. and S.J. designed the study, wrote the manuscript, researched data, contributed to the discussion, and reviewed and edited the manuscript. E.S. and A.J. researched data and reviewed and edited the manuscript. R.T.L. contributed to the discussion and reviewed and edited the manuscript. C.G.P.P. researched data and contributed to the discussion. P.M.N. contributed to the discussion and reviewed and edited the manuscript. G.R. researched data. K.M. researched data and reviewed and edited the manuscript. K.H. researched data, contributed to the discussion, and reviewed and edited the manuscript. O.M. reviewed and edited the manuscript. L.G. contributed to the discussion and reviewed and edited the manuscript. V.L. researched data and reviewed and edited the manuscript. A.M. designed the study, contributed to the discussion, and reviewed and edited the manuscript. M.O.-M. researched data and reviewed and edited the manuscript. P.R.N. designed the study, contributed to the discussion, and reviewed and edited the manuscript.

1.
Dina
C
,
Meyre
D
,
Gallina
S
, et al
.
Variation in FTO contributes to childhood obesity and severe adult obesity
.
Nat Genet
2007
;
39
:
724
726
[PubMed]
2.
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
894
[PubMed]
3.
Scott
LJ
,
Mohlke
KL
,
Bonnycastle
LL
, et al
.
A genome-wide association study of type 2 diabetes in Finns detects multiple susceptibility variants
.
Science
2007
;
316
:
1341
1345
[PubMed]
4.
Scuteri
A
,
Sanna
S
,
Chen
WM
, et al
.
Genome-wide association scan shows genetic variants in the FTO gene are associated with obesity-related traits
.
PLoS Genet
2007
;
3
:
e115
[PubMed]
5.
Fawcett
KA
,
Barroso
I
.
The genetics of obesity: FTO leads the way
.
Trends Genet
2010
;
26
:
266
274
[PubMed]
6.
Al-Attar
SA
,
Pollex
RL
,
Ban
MR
, et al
.
Association between the FTO rs9939609 polymorphism and the metabolic syndrome in a non-Caucasian multi-ethnic sample
.
Cardiovasc Diabetol
2008
;
7
:
5
[PubMed]
7.
Barber
TM
,
Bennett
AJ
,
Groves
CJ
, et al
.
Association of variants in the fat mass and obesity associated (FTO) gene with polycystic ovary syndrome
.
Diabetologia
2008
;
51
:
1153
1158
[PubMed]
8.
Freathy
RM
,
Timpson
NJ
,
Lawlor
DA
, et al
.
Common variation in the FTO gene alters diabetes-related metabolic traits to the extent expected given its effect on BMI
.
Diabetes
2008
;
57
:
1419
1426
[PubMed]
9.
Hertel
JK
,
Johansson
S
,
Raeder
H
, et al
.
Genetic analysis of recently identified type 2 diabetes loci in 1,638 unselected patients with type 2 diabetes and 1,858 control participants from a Norwegian population-based cohort (the HUNT study)
.
Diabetologia
2008
;
51
:
971
977
[PubMed]
10.
Hunt
SC
,
Stone
S
,
Xin
Y
, et al
.
Association of the FTO gene with BMI
.
Obesity (Silver Spring)
2008
;
16
:
902
904
[PubMed]
11.
Jess
T
,
Zimmermann
E
,
Kring
SI
, et al
.
Impact on weight dynamics and general growth of the common FTO rs9939609: a longitudinal Danish cohort study
.
Int J Obes (Lond)
2008
;
32
:
1388
1394
[PubMed]
12.
Marvelle
AF
,
Lange
LA
,
Qin
L
,
Adair
LS
,
Mohlke
KL
.
Association of FTO with obesity-related traits in the Cebu Longitudinal Health and Nutrition Survey (CLHNS) Cohort
.
Diabetes
2008
;
57
:
1987
1991
[PubMed]
13.
Qi
L
,
Kang
K
,
Zhang
C
, et al
.
Fat mass-and obesity-associated (FTO) gene variant is associated with obesity: longitudinal analyses in two cohort studies and functional test
.
Diabetes
2008
;
57
:
3145
3151
[PubMed]
14.
Tabara
Y
,
Osawa
H
,
Guo
H
, et al
.
Prognostic significance of FTO genotype in the development of obesity in Japanese: the J-SHIPP study
.
Int J Obes (Lond)
2009
;
33
:
1243
1248
[PubMed]
15.
Wangensteen T, Egeland T, Akselsen H, Holmen J, Undlien D, Retterstol L. FTO Genotype and Weight Gain in Obese and Normal Weight Adults From a Norwegian Population Based Cohort (the HUNT Study). Exp Clin Endocrinol Diabetes, 2010;118:649–652
16.
Jacobsson
JA
,
Risérus
U
,
Axelsson
T
,
Lannfelt
L
,
Schiöth
HB
,
Fredriksson
R
.
The common FTO variant rs9939609 is not associated with BMI in a longitudinal study on a cohort of Swedish men born 1920-1924
.
BMC Med Genet
2009
;
10
:
131
[PubMed]
17.
Luan
J
,
Kerner
B
,
Zhao
JH
, et al
.
A multilevel linear mixed model of the association between candidate genes and weight and body mass index using the Framingham longitudinal family data
.
BMC Proc
2009
;
3
(
Suppl. 7
):
S115
[PubMed]
18.
Kaakinen
M
,
Läärä
E
,
Pouta
A
, et al
.
Life-course analysis of a fat mass and obesity-associated (FTO) gene variant and body mass index in the Northern Finland Birth Cohort 1966 using structural equation modeling
.
Am J Epidemiol
2010
;
172
:
653
665
[PubMed]
19.
Holmen
J
,
Midthjell
K
,
Krüger
Ø
, et al
.
The Nord-Trøndelag Health Study 1995-97 (HUNT2): objectives, contents, methods and participation
.
Nor Epidemiol
2003
;
13
:
19
32
20.
Berglund
G
,
Elmstähl
S
,
Janzon
L
,
Larsson
SA
.
The Malmo Diet and Cancer Study. Design and feasibility
.
J Intern Med
1993
;
233
:
45
51
[PubMed]
21.
Nilsson
P
,
Berglund
G
.
Prevention of cardiovascular disease and diabetes: lessons from the Malmö Preventive Project
.
J Intern Med
2000
;
248
:
455
462
[PubMed]
22.
Purcell
S
,
Neale
B
,
Todd-Brown
K
, et al
.
PLINK: a tool set for whole-genome association and population-based linkage analyses
.
Am J Hum Genet
2007
;
81
:
559
575
[PubMed]
23.
Harris R, Bradburn M, Deeks J, et al. METAN: Stata module for fixed and random effects meta-analysis. In Statistical Software Components S456798. Chestnut Hill, MA, Boston College Department of Economics, Revised 19 May 2009
24.
Mägi
R
,
Morris
AP
.
GWAMA: software for genome-wide association meta-analysis
.
BMC Bioinformatics
2010
;
11
:
288
[PubMed]
25.
Sanghera
DK
,
Ortega
L
,
Han
S
, et al
.
Impact of nine common type 2 diabetes risk polymorphisms in Asian Indian Sikhs: PPARG2 (Pro12Ala), IGF2BP2, TCF7L2 and FTO variants confer a significant risk
.
BMC Med Genet
2008
;
9
:
59
[PubMed]
26.
Bressler
J
,
Kao
WH
,
Pankow
JS
,
Boerwinkle
E
.
Risk of type 2 diabetes and obesity is differentially associated with variation in FTO in whites and African-Americans in the ARIC study
.
PLoS ONE
2010
;
5
:
e10521
[PubMed]
27.
Yajnik
CS
,
Janipalli
CS
,
Bhaskar
S
, et al
.
FTO gene variants are strongly associated with type 2 diabetes in South Asian Indians
.
Diabetologia
2009
;
52
:
247
252
[PubMed]
28.
Berulava
T
,
Horsthemke
B
.
The obesity-associated SNPs in intron 1 of the FTO gene affect primary transcript levels
.
Eur J Hum Genet
2010
;
18
:
1054
1056
[PubMed]
29.
Zabena
C
,
González-Sánchez
JL
,
Martínez-Larrad
MT
, et al
.
The FTO obesity gene. Genotyping and gene expression analysis in morbidly obese patients
.
Obes Surg
2009
;
19
:
87
95
[PubMed]
30.
Kirkpatrick
CL
,
Marchetti
P
,
Purrello
F
, et al
.
Type 2 diabetes susceptibility gene expression in normal or diabetic sorted human alpha and beta cells: correlations with age or BMI of islet donors
.
PLoS ONE
2010
;
5
:
e11053
[PubMed]
31.
Poritsanos
NJ
,
Lew
PS
,
Mizuno
TM
.
Relationship between blood glucose levels and hepatic Fto mRNA expression in mice
.
Biochem Biophys Res Commun
2010
;
400
:
713
717
[PubMed]
32.
Bravard
A
,
Lefai
E
,
Meugnier
E
, et al
.
FTO is increased in muscle during type 2 diabetes, and its overexpression in myotubes alters insulin signaling, enhances lipogenesis and ROS production, and induces mitochondrial dysfunction
.
Diabetes
2011
;
60
:
258
268
[PubMed]
33.
Ragvin
A
,
Moro
E
,
Fredman
D
, et al
.
Long-range gene regulation links genomic type 2 diabetes and obesity risk regions to HHEX, SOX4, and IRX3
.
Proc Natl Acad Sci USA
2010
;
107
:
775
780
[PubMed]
34.
López-Bermejo
A
,
Petry
CJ
,
Díaz
M
, et al
.
The association between the FTO gene and fat mass in humans develops by the postnatal age of two weeks
.
J Clin Endocrinol Metab
2008
;
93
:
1501
1505
[PubMed]
35.
Hakanen
M
,
Raitakari
OT
,
Lehtimäki
T
, et al
.
FTO genotype is associated with body mass index after the age of seven years but not with energy intake or leisure-time physical activity
.
J Clin Endocrinol Metab
2009
;
94
:
1281
1287
[PubMed]
36.
Rzehak
P
,
Scherag
A
,
Grallert
H
, et al
;
GINI and LISA Study Group
.
Associations between BMI and the FTO gene are age dependent: results from the GINI and LISA birth cohort studies up to age 6 years
.
Obes Facts
2010
;
3
:
173
180
[PubMed]
37.
Sonestedt
E
,
Roos
C
,
Gullberg
B
,
Ericson
U
,
Wirfält
E
,
Orho-Melander
M
.
Fat and carbohydrate intake modify the association between genetic variation in the FTO genotype and obesity
.
Am J Clin Nutr
2009
;
90
:
1418
1425
[PubMed]
38.
Vimaleswaran
KS
,
Li
S
,
Zhao
JH
, et al
.
Physical activity attenuates the body mass index-increasing influence of genetic variation in the FTO gene
.
Am J Clin Nutr
2009
;
90
:
425
428
[PubMed]
39.
Andreasen
CH
,
Stender-Petersen
KL
,
Mogensen
MS
, et al
.
Low physical activity accentuates the effect of the FTO rs9939609 polymorphism on body fat accumulation
.
Diabetes
2008
;
57
:
95
101
[PubMed]

Supplementary data