Background: The etiology of hormonally related cancers, such as breast and colon, has been linked to hyperinsulinemia and insulin resistance, the insulin-like growth factor (IGF) axis, and obesity.

Methods: Data from 57 women (ages 30-50 years) were used to observationally examine cross-sectional and longitudinal relations between body fat (from dual-energy X-ray absorptiometry), insulin, IGF-I, and IGF-binding proteins (IGFBP-1, IGFBP-2, and IGFBP-3).

Results: At baseline, participants who had greater than median body fat and insulin levels, >39% and >4.5 microunits/mL, respectively, had 2.3- to 2.6-fold lower IGFBP-1 (P < 0.004) and 1.9- to 2.0-fold lower IGFBP-2 (P < 0.004) compared with other participants; IGF-I and IGFBP-3 levels did not differ by body fat or insulin levels. Over 39 weeks, a 1 microunit/mL reduction in fasting insulin was associated with a 17% increase in IGFBP-1 (P = 0.02) and a 24% increase in IGFBP-2 (P = 0.02) compared with participants who did not reduce insulin; 2.0% loss of body fat over time did not alter IGFBP-1 or IGFBP-2 levels after adjustment for insulin. IGF-I and IGFBP-3 did not change in participants who lost body fat percentage or insulin over time.

Conclusions: These observational associations are consistent with the hypothesis that elevated insulin and body fat are associated with decreased IGFBP-1 and IGFBP-2 levels cross-sectionally; they further imply that IGFBP-1 and IGFBP-2 levels may be altered through change in insulin over time. By contrast, no cross-sectional or longitudinal associations were noted between IGF-I and IGFBP-3 with insulin or body fat. (Cancer Epidemiol Biomarkers Prev 2007;16(3):593–7)

Incidence of several of the most common hormonally related cancers, breast, colorectal, and perhaps prostate and endometrial, has been linked to obesity (1-5), insulin (6-9), and the insulin-like growth factor (IGF) axis (10-15). These potential cancer biomarkers are intriguing for cancer control purposes because they may interact and be modifiable through behavioral or pharmaceutical interventions.

Recent evidence from population-based studies suggests that obesity, insulin, and the IGF axis may interact to increase cancer risk (9, 16-18). Higher levels of adipose, including visceral fat, are associated with increased blood levels of free fatty acids and serum insulin (19). Insulin independently leads to cellular growth and proliferation and has been associated with several cancers (8). The interactions between body fat and insulin with the IGF axis are less well understood, which poses a challenge in developing cancer control interventions. In extracellular fluids, IGF-I is complexed to six distinct IGF-binding proteins (IGFBP-1, IGFBP-2, IGFBP-3, IGFBP-4, IGFBP-5, and IGFBP-6), which function in part to bind >95% of circulatory IGF-I and thereby modulate the availability of free IGF-I to act at its receptors (13). Circulating IGF-I and IGFBP-3 levels are determined by liver production, which is regulated by growth hormone activity and nutritional status (20). IGFBP-3 and IGF-I have long half-lives in circulation; as they are growth hormone dependent, they represent a stable index of long-term change in growth hormone availability and action (21). As reviewed by Lukanova et al. (22), with increased adiposity, physiologic growth hormone secretion is impaired and growth hormone responses to all stimuli are decreased; conversely, insulin enhances growth hormone–stimulated synthesis of IGF-I and IGFBP-3 through up-regulation of growth hormone receptors (23). Therefore, behavioral and pharmaceutical interventions may affect IGF-I and its binding proteins differentially according to factors that influence both growth hormone and nutritional status, including insulin and body composition (20). Insulin may affect IGF-I levels and activity by contributing to growth hormone stimulation of IGF-I production (13) and by increasing IGF-I bioavailability through direct inhibition of IGFBP-1 and IGFBP-2 synthesis at the liver (24-30). Associations of body fat and insulin with cancer risk may be mediated through alterations of bioavailable levels of IGF-I. Indeed, IGFBP-1 and IGFBP-2 have been associated negatively in some (14, 15), although not all (31-34), studies with cancer development.

Prior cross-sectional studies examining associations between body composition, insulin, and the IGF axis have generally used crude measures of body fatness, such as body mass index (BMI). We have dual-energy X-ray absorptiometry body composition data, in addition to fasting insulin and IGF axis data on 57 women at multiple time points over 39 weeks. We used these data to test two hypotheses. The first is that elevated body fat and fasting insulin are associated with lower levels of IGFBP-1 and IGFBP-2 but not with IGF-I or IGFBP-3. The second hypothesis is that decreases in insulin or in body fat levels are associated with increases in IGFBP-1 and IGFBP-2 levels over 39 weeks. The data for these analyses are from a previously completed randomized controlled trial of twice-weekly weight training in 57 women. We have previously reported a nonsignificant intervention effect on insulin change over 39 weeks in this group of women (35); however, there were an equal number of participants in both the intervention treatment and control groups who decreased their insulin level over 39 weeks. Therefore, to explore the associations between body fat, insulin, and the IGF axis, we examined the 57 participants as one group, controlling for assignment to the treatment condition, cross-sectionally at baseline and longitudinally over 39 weeks.

The weight gain prevention study was a 39-week randomized-controlled trial of twice-weekly weight training in 60 women (n = 30 intervention, n = 30 control). Details of participant recruitment and inclusion criteria have been described in detail (35, 36). Briefly, during December 1999 and January 2000, 60 women ages 30 to 50 years were recruited from female faculty, staff, and students at the University of Minnesota (Minneapolis, MN). Participants were randomized in a blinded fashion, stratified by decade of age (30-39 versus 40-50). The intervention design has been described (35, 36). Briefly, all participants were asked to avoid altering their baseline low-intensity aerobic or stretching exercises (most commonly one to two weekly walks) and dietary habits for the purpose of weight change for the study duration. The treatment group was enrolled in a 50-min strength training class held twice weekly for 15 weeks at the University of Minnesota Recreation Center. Participants did three sets each of nine common strength training exercises, lifting as much weight as they could for 8 to 10 repetitions per set. Through the first 15 weeks of the intervention, participants progressively increased the weight of a given exercise as described previously (35, 36). After 15 weeks, treatment group participants continued to exercise unsupervised for 6 months at the same exercise facility. Participants maintained exercise session logs, in which they recorded information about number of sets and the weight load for each exercise done per strength training session. The control group participated in measurements only. During the 15-week supervised intervention and the following 6 months of unsupervised exercise, 92% and 83% of prescribed exercise sessions were completed, respectively. Strength changes from baseline to 15 and 39 weeks corroborated self-reported adherence data as the treatment group showed increases of 19% and 24% for bench press and 13% and 20% for leg press over 15 and 39 weeks, respectively (P < 0.0001 for bench press, P < 0.004 for leg press when compared with the control group; refs. 35, 36). The main results of the intervention have been reported (35, 36). The study protocol was reviewed by and followed all regulations of the University of Minnesota Institutional Review Board for the protection of human subjects in research.

Dropouts and loss to follow-up: there were incomplete baseline blood samples for two participants (one each in the treatment and control groups); one treatment group participant was diagnosed with Grave's disease before study completion. Data for these three participants are excluded from analyses; baseline analyses include the remaining 57 women. Three women dropped out of the study after baseline (one treatment and two control group participants); longitudinal analyses include the 54 women who completed the study.

Measurements

Measurements were taken at baseline and 39 weeks later (study completion) at a measurement visit at the University of Minnesota General Clinical Research Center. Participants refrained from moderate intensity physical activity for 48 h before measurements. Body weight and height, body composition, and blood draws were done by clinical research nurses blinded to treatment group status between 6:30 and 9:30 a.m., after a 12-h fast, and between 5 and 11 days after the start of menstrual flow for menstruating participants. Participants wore hospital gowns for measurements.

Serum insulin levels were measured by a two-site immunoenzymatic assay at the Mayo Clinic (Rochester, MN) as described previously (33). ELISAs of IGF-I, IGFBP-1, IGFBP-2, and IGFBP-3 were done at Diagnostic Systems Laboratories (Webster, TX) as described previously (36). All assays were done in duplicate, with the mean used for analysis, by technicians blinded to treatment status. For IGFBP-2, measurements from a single participant ID were done in a single assay using the same reagent batch. For all other assays, technicians matched reagent batch to measurement time point for all participant ID numbers.

Weight was measured on a digital scale and height was assessed using a mounted stadiometer, both calibrated daily (Scale-Tronix 5005 stand-on digital scale, Scale-Tronix, White Plains, NY). Waist circumference, at the level of the umbilicus, was measured in duplicate; the mean was used for analysis. Body composition was measured on the Lunar Prodigy Dual X-ray Absorptiometer software version 2.15 (Lunar Radiation Corp., Madison, WI), calibrated monthly with daily calibration checks. Body fat percentage was calculated as the proportion of non–bone tissue that was fat.

Participants completed three surveys: the Typical Weekly Physical Activity Survey (37), 4-day food records, and a demographic survey (given at the baseline measurement visit only) as described previously (35, 36).

Statistical Analysis

Statistical Analysis System version 8.2 (SAS, Cary, NC) was used for all analyses. Baseline demographic, physical activity energy expenditure, and total energy intake data were compared in the entire cohort, across baseline body fat percentage (above and below median, 39.1%), and across baseline insulin level (above and below median, 4.5 microunits/mL) using two-sided t tests for continuous and χ2 tests for categorical variables.

Least-squares means of IGF axis proteins and body composition variables were compared across four groups of the participants. These groups were determined by baseline body fat percentage (above and below median, 39.1%) and by baseline insulin level (above and below median, 4.5 microunits/mL): group 1 (n = 20), ≥39.1% body fat and ≥4.5 microunits/mL insulin; group 2 (n = 9), ≥39.1% body fat and <4.5 microunits/mL insulin; group 3 (n = 9), <39.1% body fat and ≥4.5 microunits/mL insulin; and group 4 (n = 19), <39.1% body fat and <4.5 microunits/mL insulin. All ANOVAs used generalized linear models (Proc GLM). Potential confounders were tested, including treatment group status, baseline energy consumption and expenditure, menopausal status, and age; these variables did not alter the interpretation of the results and were therefore not included in final models.

To examine longitudinal change in mean levels of IGF axis proteins, participants were divided into groups for two separate analyses based on the following cut points: (a) the first analyses divided all participants into two groups based on those who experienced <1 or ≥1 microunit/mL loss of insulin over 39 weeks and (b) the second analyses divided all participants into two groups based on those who experienced <2.0% or ≥2.0% loss of body fat over 39 weeks. These cut points were chosen because both 2.0% loss of body fat and 1 microunit/mL insulin have been reported as levels of change attainable from exercise interventions (38, 39). In the Weight Gain Prevention Study (WGPS), there was a lack of effect of the intervention on insulin; specifically, 10 women in each intervention group decreased their insulin levels over the 39-week intervention (36). Therefore, in these analyses, we examined insulin effects independent of the intervention and included treatment group status as a covariate in the models. Additional adjustment for baseline levels of the dependent variable (insulin, IGF-I, and IGFBP-1, IGFBP-2, and IGFBP-3), hormone replacement therapy or hormonal contraception use, demographic variables, age, physical activity (baseline or change), and energy intake (baseline or change) were tested but not included in final models as they did not alter the results presented. Change in insulin as a continuous variable was added to the models to examine whether insulin mediated the effect of change in body fat percentage on IGF axis proteins.

All statistical tests and corresponding P values were two sided; a P value of <0.05 is reported as statistically significant.

Participants in the Weight Gain Prevention Study (WGPS) were mostly married, college educated, and Caucasian as described previously (36). There were no differences in age, race, marital status, energy intake, or energy expenditure in participants who were above versus below median body fat percentage (data not shown); by contrast, women who had <39% body fat were more likely to be college educated compared with women who had ≥39% body fat (96% versus 76%). There were no differences in age, race, marital status, education level, or energy expenditure in participants who were above median insulin level versus below; by contrast, women who had ≥4.5 microunits/mL insulin had greater energy intake at baseline compared with those with <4.5 microunits/mL insulin (2,024.3 ± 71.8 versus 1,645.0 ± 72.8, respectively).

At baseline, women (n = 20) who were above median body fat percentage (≥39%) and insulin level (≥4.5 microunits/mL) had 2.3- to 2.6-fold lowered IGFBP-1 (P < 0.004) and 1.8- to 2.0-fold lowered IGFBP-2 (P < 0.004) levels, respectively, compared with other participants (Table 1). IGFBP-1 and IGFBP-2 levels did not differ significantly among the other three groups. There was also a significant difference in IGFBP-1 and IGFBP-2 levels between those with higher and lower insulin among the subgroup with higher body fat, whereas there was no difference according to insulin in those with lower body fat. There were no associations between body fat percentage (≥39 versus <39), fasting insulin (≥4.5 microunits/mL versus <4.5 microunits/mL), or the interaction of the two with mean levels of IGF-I or IGFBP-3. There was a 2.8-fold difference in fasting insulin level over the four groups despite a study participation eligibility restriction of normal fasting insulin at baseline (≤10 microunits/mL; ref. 36). Body fat percentage varied 1.5-fold, although there was a nonstatistically significant 8% variation in lean mass over the groups.

Table 1.

Baseline cross-sectional levels of anthropometric and hormonal values by baseline insulin and body fat percentage (mean ± SD)

≥4.5 microunits/mL<4.5 microunits/mL
IGF-I (mg/dL)   
    ≥39% body fat 222.6 ± 54.0 218.1 ± 51.7 
    <39% body fat 238.7 ± 66.1 233.5 ± 63.5 
IGFBP-1 (mg/dL)   
    ≥39% body fat 26.0 ± 16.5 a 60.4 ± 26.5 b 
    <39% body fat 61.5 ± 40.4 b 66.4 ± 29.8 b 
IGFBP-2 (mg/dL)   
    ≥39% body fat 334.8 ± 191.3 a 625.3 ± 192.2 b 
    <39% body fat 685.2 ± 352.0 b 636.7 ± 247.2 b 
IGFBP-3 (mg/dL)   
    ≥39% body fat 3,670.2 ± 424.4 3,534.2 ± 249.1 
    <39% body fat 3,518.3 ± 860.6 3,718.4 ± 805.9 
Insulin   
    ≥39% body fat 7.6 ± 3.1 a 3.1 ± 0.6 b 
    <39% body fat 5.4 ± 1.2 a 2.8 ± 0.9 b 
Body fat %   
    ≥39% body fat 45.8 ± 3.6 a 44.2 ± 5.0 a 
    <39% body fat 35.4 ± 4.7 b 33.6 ± 3.4 b 
Lean mass   
    ≥39% body fat 42.9 ± 5.2 a 39.9 ± 5.7 ab 
    <39% body fat 39.9 ± 5.6 ab 39.5 ± 3.7 b 
≥4.5 microunits/mL<4.5 microunits/mL
IGF-I (mg/dL)   
    ≥39% body fat 222.6 ± 54.0 218.1 ± 51.7 
    <39% body fat 238.7 ± 66.1 233.5 ± 63.5 
IGFBP-1 (mg/dL)   
    ≥39% body fat 26.0 ± 16.5 a 60.4 ± 26.5 b 
    <39% body fat 61.5 ± 40.4 b 66.4 ± 29.8 b 
IGFBP-2 (mg/dL)   
    ≥39% body fat 334.8 ± 191.3 a 625.3 ± 192.2 b 
    <39% body fat 685.2 ± 352.0 b 636.7 ± 247.2 b 
IGFBP-3 (mg/dL)   
    ≥39% body fat 3,670.2 ± 424.4 3,534.2 ± 249.1 
    <39% body fat 3,518.3 ± 860.6 3,718.4 ± 805.9 
Insulin   
    ≥39% body fat 7.6 ± 3.1 a 3.1 ± 0.6 b 
    <39% body fat 5.4 ± 1.2 a 2.8 ± 0.9 b 
Body fat %   
    ≥39% body fat 45.8 ± 3.6 a 44.2 ± 5.0 a 
    <39% body fat 35.4 ± 4.7 b 33.6 ± 3.4 b 
Lean mass   
    ≥39% body fat 42.9 ± 5.2 a 39.9 ± 5.7 ab 
    <39% body fat 39.9 ± 5.6 ab 39.5 ± 3.7 b 

NOTE: n = 20: ≥4.5 microunits/mL and ≥39% body fat; n = 9: ≥4.5 microunits/mL and <39% body fat; n = 9: <4.5 microunits/mL and ≥39% body fat; n = 19: <4.5 microunits/mL and <39% body fat. ANOVA tests: statistically significant different values have different letters; for example, value a is significantly different than value b.

Participants who lost ≥1.0 microunit/mL insulin from baseline to 39 weeks, in comparison with those who did not, had increased IGFBP-1 and IGFBP-2 levels by 17% (P = 0.02) and 24% (P = 0.02) at 39 weeks, respectively (Table 2). By contrast, loss of insulin over 39 weeks was not associated with significant change in levels of either IGF-I or IGFBP-3. We also examined whether a change in body fat percentage over 39 weeks (participants who lost ≥2.0% versus those who had <2.0% body fat change) would be associated with alterations of IGF axis proteins (data not shown). For these analyses, we observed IGFBP-1 and IGFBP-2 increased by 9% and 11% (not statistically significant), respectively, and there was no difference in IGF-I or IGFBP-3 level from 0 to 39 weeks in the group that lost ≥2.0% versus those who had <2.0% body fat change by 39 weeks. The group that lost ≥2.0% body fat experienced a 1.1 microunit/mL decrease in insulin over this time period. The addition of change in insulin over 39 weeks to the regression models attenuated the association between change in body fat on IGFBP-2.

Table 2.

Change in IGF axis proteins by change in insulin over 39 wks (mean ± SD)

Insulin loss
P
≥1 microunit/mL (n = 10)<1 microunit/mL (n = 44)
Insulin −2.2 ± 1.02 1.2 ± 1.9 <0.0001 
IGF-I 24.9 ± 51.0 20.0 ± 56.7 0.80 
IGFBP-1 7.5 ± 30.8 −11.2 ± 19.8 0.02 
IGFBP-2 113.3 ± 219.5 −15.2 ± 138.4 0.02 
IGFBP-3 112.1 ± 901.5 371.3 ± 940.5 0.43 
Insulin loss
P
≥1 microunit/mL (n = 10)<1 microunit/mL (n = 44)
Insulin −2.2 ± 1.02 1.2 ± 1.9 <0.0001 
IGF-I 24.9 ± 51.0 20.0 ± 56.7 0.80 
IGFBP-1 7.5 ± 30.8 −11.2 ± 19.8 0.02 
IGFBP-2 113.3 ± 219.5 −15.2 ± 138.4 0.02 
IGFBP-3 112.1 ± 901.5 371.3 ± 940.5 0.43 

NOTE: Least-squares means from general linear regression models adjusted for treatment group.

In this sample of 30- to 50-year-old women, there were inverse cross-sectional and longitudinal associations of insulin and body fat percentage with IGFBP-1 and IGFBP-2 but not with IGF-I or IGFBP-3. Change in IGFBP-1 and IGFBP-2 levels alters bioavailable IGF-I levels, which may then alter cancer risk. Although studies examining associations between IGFBP-1 and IGFBP-2 and cancer risk have been mixed (14, 15, 31-34), the findings from these analyses may have relevance for understanding mechanisms through which change in insulin levels and or body composition may alter cancer risk.

The results of the present cross-sectional analyses complement the findings of other studies that have examined cross-sectional associations of body fat (generally BMI) with insulin and the IGF axis. Similar to the present findings, several authors have reported that total IGF-I and IGFBP-3 are not correlated with BMI, whereas IGFBP-1 and/or IGFBP-2 are inversely correlated with BMI (r = −0.3 to −0.5 for IGFBP-1 and r = −0.3 to −0.5 for IGFBP-2; refs. 14, 21, 40-45), insulin, or C-peptide (r = −0.21 to −0.46 for IGFBP-1 and r = −0.16 to −0.26 for IGFBP-2; refs. 40, 42, 46-48). In a cross-sectional study of 400 women, Lukanova et al. (42) reported that diabetics (mean fasting insulin, 20.81 microunits/mL) had 2.4-fold lower IGFBP-1 and 1.8-fold lower IGFBP-2 compared with nondiabetics (mean fasting insulin, 8.59 microunits/mL). In contrast to our findings, some studies have observed decreased (49, 50) or increased (51, 52) total IGF-I levels in obese compared with nonobese participants. Whereas several studies did not examine associations between IGF-I and insulin, others have shown a positive correlation between total IGF-I and insulin (21). Additionally, nonlinear relations of IGF-I have been reported with BMI in one study of women (53) and with BMI and insulin in men but not women in another study (22). Whereas we did not observe an association between elevated body fat and IGFBP-3, Frystyk et al. (47) reported that obese nondiabetic subjects had elevated IGFBP-3 compared with lean subjects. There is a diversity in the findings about any association of body fat or insulin with IGFBP-3 (22, 41, 42, 44, 45, 47, 49-51, 54, 55). These collective findings indicate that body fat and insulin may be associated inversely with IGFBP-1 and IGFBP-2 but are less likely to be associated with total IGF-I levels and may or may not be associated with IGFBP-3.

Interactions between obesity, insulin, and the IGF axis are complex and incompletely understood. As noted above, although several authors have reported altered levels of IGFBP-1 and IGFBP-2 by body fat and/or insulin, observations have been mixed about levels of IGF-I or IGFBP-3 by body fat and/or insulin. The varied observations may be attributable in part to differences in age, sex, degree of obesity or elevated insulin, and nutritional factors in study participants. Additionally, a nonlinear relation between total IGF-I and BMI (22, 53) or insulin (22) has been observed, indicating that other factors may play a role in the regulation of IGF-I in obesity. Age is a predictor of IGFBP-3 (42); we may not have seen variation in IGFBP-3 in our study in part because participants were close in age. Obesity is associated with growth hormone hyposecretion; therefore, it may be expected that total IGF-I concentrations would be lower in obese compared with nonobese subjects (21). However, several of the studies reviewed above did not report lower IGF-I in obese compared with nonobese subjects. Obesity is also associated with hyperinsulinemia, which may increase hepatic IGF-I production, as reviewed above (22). It is also possible that the obesity and hyperinsulinemia may be associated with decreased total IGF-I levels but with increased free IGF-I levels. As reviewed above, insulin regulates IGFBP-1 and IGFBP-2 levels through inhibition of hepatic synthesis. IGFBP-1 and IGFBP-2 are regulators of IGF-I bioavailability (13); when IGFBP-1 and IGFBP-2 levels are decreased, free IGF-I levels increase (22). Our findings are consistent with the theory that insulin exerts its effects on IGF-I through alterations in IGFBP-1 and IGFBP-2. We found that relatively small cross-sectional or longitudinal variations in insulin levels altered levels of these binding proteins. The women in this study were all normoinsulinemic at baseline, which may explain why we did not observe associations cross-sectionally or longitudinally between IGF-I and IGFBP-3 if differences in insulin levels were not great enough to affect levels of these proteins. The present analyses included models, which adjusted for factors that may be associated with alterations of IGF axis variables, such as tobacco use, age, diet, and estrogens (50). However, we did not measure free IGF-I or growth hormone in this study and therefore cannot fully determine relations between body fat, insulin, and free IGF-I. That said, the observational studies that have reported associations of IGF-I and hormonally related cancers have generally used total IGF-I as the biomarker of interest, not free IGF-I (10-12, 14, 15, 32-34).

Two prior studies have examined specifically associations between change in body size or insulin, secondary to an intervention, with change in IGF axis proteins, including IGFBP-1 and IGFBP-2, over time. Hellenius et al. (1995) conducted a randomized controlled trial in 157 normoglycemic 35- to 60-year-old men to examine whether a 6-month intervention of diet, exercise, or diet and exercise combined compared with a control group would alter levels of insulin, IGF-I, and IGFBP-1 (37). At 6 months, fasting insulin levels decreased 16% to 19%, IGFBP-1 increased 15% to 26%, and IGF-I levels did not change in the intervention groups compared with control group. Increased IGFBP-1 levels at 6 months were positively correlated with decreased BMI (r = 0.30, P < 0.01), waist-to-hip ratio (r = 0.30, P < 0.001), and increased exercise (r = 0.23, P < 0.05; ref. 37). The Diet and Androgens study was a 5-month randomized controlled dietary intervention designed to alter body composition and hormonal levels in 50- to 65-year-old postmenopausal women (49 treatment, 50 control; ref. 46). The authors reported significant reductions in BMI (6%), waist circumference (5%), and C-peptide (19%), a significant increase in IGFBP-1 (12.1%) and IGFBP-2 (30.4%), and no change in IGF-I or IGFBP-3 in the intervention compared with control group (46). There was a nonsignificant 10.4% decrease compared with 5.1% increase in fasting insulin in the intervention compared with control group, respectively (46). Longitudinally, changes in IGFBP-1 and IGFBP-2 levels were negatively correlated with change in BMI in the intervention group (r = −0.34 and −0.17, respectively), whereas insulin was positively correlated with IGFBP-2 (r = 0.20) but not IGFBP-1 (46). The longitudinal findings of these two studies and the current findings collectively imply that change in insulin and possibly body fat may alter IGFBP-1 and IGFBP-2 over time.

The limitations of these analyses require discussion. This report is a post hoc analysis of a completed study. The availability of higher quality body fat measures compared with prior studies made it worthwhile to explore associations in this data set. Although we adjusted for treatment group status for the longitudinal analyses, we did not find evidence of confounding by this variable. However, it is possible that a residual effect of the intervention, or individual change in insulin, body fat, or IGF-I, may have altered the observed associations beyond what was observable statistically. For IGF-I, IGFBP-1, and IGFBP-3, baseline and follow-up assays were measured with different reagent batches; batch-to-batch variability in assays could obscure biological changes. Self-report instruments were used to measure changes in energy expenditure and intake. The limitations of such self-reports are well established.

In summary, we found that the combination of higher body fat percentage and insulin was associated with lowered levels of IGFBP-1 and IGFBP-2 at baseline and that decreased insulin levels were associated with increased IGFBP-1 and IGFBP-2 over 39 weeks. High levels of IGF-I have been associated with increased cancer risk. Because IGF-I circulates bound to high-affinity IGFBPs, changes in IGFBP levels likely affect the delivery of IGF-I to target tissues. Lowered insulin levels and increased IGFBP-1 and IGFBP-2 levels could limit the activation of the insulin and type I IGF receptor in target tissues by reducing the delivery of ligand. The interactions of these variables may play a role in the mechanisms between the observed associations between insulin resistance and hormonally related cancers.

Grant support: Minnesota Obesity Center Pilot and Feasibility, NIH grant DK50456, University of Minnesota General Clinical Research Center grant M01-RR00400, Public Health Service Cancer Center Support grant P30 CA77398, and Division of Epidemiology student project grant. R.L. Ahmed was supported by NIH grants T32 GM08244-15 and T32 CA09607-15, University of Minnesota Thomas Shevlin predoctoral fellowship, and National PEO Foundation scholarship.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank the participants for their efforts.

1
Albanes D. Energy balance, body size, and cancer.
Crit Rev Oncol Hematol
1990
;
10
:
283
–303.
2
Folsom AR, Kaye SA, Prineas RJ, Potter JD, Gapstur SM, Wallace RB. Increased incidence of carcinoma of the breast associated with abdominal adiposity in postmenopausal women.
Am J Epidemiol
1990
;
131
:
794
–803.
3
Folsom AR, Kushi LH, Anderson KE, et al. Associations of general and abdominal obesity with multiple health outcomes in older women: the Iowa Women's Health Study.
Arch Intern Med
2000
;
160
:
2117
–28.
4
Russo A, Franceschi S, La Vecchia C, et al. Body size and colorectal-cancer risk.
Int J Cancer
1998
;
78
:
161
–5.
5
Schapira DV, Kumar NB, Lyman GH, et al. Abdominal obesity and breast cancer risk.
Ann Intern Med
1990
;
112
:
182
–6.
6
Giovannucci E. Insulin and colon cancer.
Cancer Causes Control
1995
;
6
:
164
–79.
7
Kaaks R. Nutrition, hormones, and breast cancer: is insulin the missing link?
Cancer Causes Control
1996
;
7
:
605
–25.
8
Gupta K, Krishnaswamy G, Karnad A, Peiris AN. Insulin: a novel factor in carcinogenesis.
Am J Med Sci
2002
;
323
:
140
–5.
9
Chang CK, Ulrich CM. Hyperinsulinaemia and hyperglycaemia: possible risk factors of colorectal cancer among diabetic patients.
Diabetologica
2003
;
46
:
595
–607.
10
Giovannucci E. Insulin, insulin-like growth factors, and colon cancer: a review of the evidence.
J Nutr
2001
;
131
:
3109
–20S.
11
Sachdev D, Yee D. The IGF system and breast cancer.
Endocr Relat Cancer
2001
;
8
:
197
–209.
12
Sugumar A, Liu YC, Xia Q, Koh YS, Matsuo K. Insulin-like growth factor (IGF)-I and IGF-binding protein 3 and the risk of premenopausal breast cancer: a meta-analysis of the literature.
Int J Cancer
2004
;
111
:
293
–7.
13
Jones JI, Clemmons DR. Insulin-like growth factors and their binding proteins: biological actions.
Endocr Rev
1995
;
16
:
3
–34.
14
Kaaks R, Toniolo P, Akhmedkhanov A, et al. Serum C-peptide, insulin-like growth factor (IGF)-I, IGF-binding proteins, and colorectal cancer risk in women.
J Natl Cancer Inst
2000
;
92
:
1592
–600.
15
Lukanova A, Zeleniuch-Jacquotte A, Lundin E, et al. Prediagnostic levels of C-peptide, IGF-I, IGFBP-1, -2, and -3, and risk of endometrial cancer.
Int J Cancer
2004
;
108
:
262
–8.
16
Trevisan M, Liu J, Muti P, Misciagna G, Menotti A, Fucci F; Risk Factors and Life Expectancy Research Group. Markers of insulin resistance and colorectal cancer mortality.
Cancer Epidemiol Biomarkers Prev
2001
;
10
:
937
–41.
17
Colangelo LA, Gapstur SM, Gann PH, Dyer AR, Liu K. Colorectal cancer mortality and factors related to the insulin resistance syndrome.
Cancer Epidemiol Biomarkers Prev
2002
;
11
:
385
–91.
18
Ahmed RL, Schmitz KH, Anderson KE, Rosamond WD, Folsom AR. The metabolic syndrome and risk of incident colorectal cancer.
Cancer
2006
;
107
:
28
–36.
19
Brochu M, Starling RD, Tchernof A, Matthews DE, Garcia-Rubi E, Poehlman ET. Visceral adipose tissue is an independent correlate of glucose disposal in older obese postmenopausal women.
J Clin Endocrinol Metab
2000
;
85
:
2378
–84.
20
Nemet D, Connolly PH, Pontello-Pescatello AM, et al. Negative energy balance plays a major role in the IGF-I response to exercise training.
J Appl Physiol
2004
;
96
:
276
–82.
21
Nam SY, Lee EJ, Kim KR, et al. Effect of obesity on total and free insulin-like growth factor (IGF)-1, and their relationship to IGF-binding protein (BP)-1, IGFBP-2, IGFBP-3, insulin, and growth hormone.
Int J Obes Relat Metab Disord
1997
;
21
:
355
–9.
22
Lukanova A, Soderberg S, Stattin P, et al. Nonlinear relationship of insulin-like growth factor (IGF)-1 and IGF-1/IGF-binding protein-3 ratio with indices of adiposity and plasma insulin concentrations (Sweden).
Cancer Causes Control
2002
;
13
:
509
–16.
23
Baxter RC, Tuttle JR. Regulation of hepatic growth hormone receptors by insulin.
Biochem Biophys Res Commun
1978
;
84
:
350
–7.
24
Underwood LE, Thissen JP, Lemozy JP, Ketelslegers JM, Clemmons DR. Hormonal and nutritional regulation of IGF-I and its binding proteins.
Horm Res
1994
;
42
:
145
–51.
25
Clemmons DR. Insulin-like growth factor binding proteins and their role in controlling IGF actions.
Cytokine Growth Factor Rev
1997
;
8
:
45
–62.
26
Bereket A, Lang CH, Biethen SL, Wilson TA. Insulin-like growth factor-binding protein-2 and insulin: studies in children with type 1 diabetes mellitus and maturity-onset diabetes of the young.
J Clin Endocrinol Metab
1995
;
80
:
3647
–52.
27
Clemmons DR, Snyder DK, Busby WHJ. Variables controlling the secretion of insulin-like growth factor binding protein-2 in normal human subjects.
J Clin Endocrinol Metab
1991
;
73
:
727
–33.
28
Brismar K, Hall K. Clinical applications of IGFBP-1 and its regulation.
Growth Regul
1993
;
3
:
98
–100.
29
Suikkari AM, Koivisto VA, Rutanen EM, Yki-Jarvinen H, Karonen SL, Seppala M. Insulin regulates the serum levels of low molecular weight insulin-like growth factor-binding protein.
J Clin Endocrinol Metab
1988
;
66
:
266
–72.
30
Boni-Schnetzler M, Schmid C, Mary JL, et al. Insulin regulates the expression of the insulin-like growth factor binding protein 2 mRNA in rat hepatocytes.
Mol Endocrinol
1990
;
4
:
1320
–6.
31
Palmqvist R, Stattin P, Rinaldi S, et al. Plasma insulin, IGF-binding proteins-1 and -2, and risk of colorectal cancer: a prospective study in northern Sweden.
Int J Cancer
2003
;
107
:
89
–93.
32
Stattin P, Bylund A, Rinaldi S, et al. Plasma insulin-like growth factor-I, insulin-like growth factor-binding proteins, and prostate cancer risk: a prospective study.
J Natl Cancer Inst
2000
;
92
:
1910
–7.
33
Kaaks R, Lundin E, Manjer J, et al. Prospective study of IGF-I, IGF-binding proteins, and breast cancer risk, in northern and southern Sweden.
Cancer Causes Control
2002
;
13
:
307
–16.
34
Keinan-Boker L, de Mesquita HBB, Kaaks R, et al. Circulating levels of insulin-like growth factor I, its binding proteins-1, -2, and -3, C-peptide, and risk of postmenopausal cancer.
Int J Cancer
2003
;
106
:
90
–5.
35
Schmitz KH, Ahmed RL, Yee D. Effects of a 9-month strength training intervention on insulin, insulin-like growth factor (IGF)-I, IGF-binding protein (IGFBP)-1, and IGFBP-3 in 30-50-year-old women.
Cancer Epidemiol Biomarkers Prev
2002
;
11
:
1597
–604.
36
Schmitz KH. Jensen MD. Kugler KC. Jeffery RW. Leon AS. Strength training for obesity prevention in midlife women.
Int J Obes Relat Metab Disord
2003
;
27
:
326
–33.
37
Ainsworth BE, Haskell WL, Whitt MC. Compendium of physical activities: an update of activity codes and MET intensities.
Med Sci Sports Exerc
2000
;
32
:
S498
–504.
38
Houmard JA, Tanner CJ, Slentz CA, Duscha BD, McCartney JS, Kraus WE. Effect of the volume and intensity of exercise training on insulin sensitivity.
J Appl Physiol
2004
;
96
:
101
–6.
39
McTiernan A, Tworoger SS, Rajan KB, et al. Effect of exercise on serum androgens in postmenopausal women: a 12-month randomized clinical trial.
Cancer Epidemiol Biomarkers Prev
2004
;
13
:
1099
–105.
40
Hellenius MB, Brismar KE, Berglund BH, De Faire UH. Effect on glucose tolerance, insulin secretion, insulin-like growth factor-1 and its binding protein, IGFBP-1, in a randomized controlled diet and exercise study in healthy, middle-aged men.
J Intern Med
1995
;
238
:
121
–30.
41
Deuschle M, Blum WF, Frystyk J, et al. Endurance training and its effect upon the activity of the GH-IGFs system in the elderly.
Int J Sports Med
1998
;
19
:
250
–4.
42
Lukanova A, Toniolo P, Akhmedkhanov A, et al. A cross-sectional study of IGF-I determinants in women.
Eur J Cancer Prev
2001
;
10
:
443
–52.
43
Voskuil DW, de Mesquita HB, Kaaks R, et al. Determinants of circulating insulin-like growth factor (IGF)-I and IGF binding proteins 1-3 in premenopausal women: physical activity and anthropometry (Netherlands).
Cancer Causes Control
2001
;
12
:
951
–8.
44
Allen NE, Appleby PN, Kaaks R, Rinaldi S, Davey GK, Key TJ. Lifestyle determinants of serum insulin-like growth-factor-I (IGF-I), C-peptide, and hormone binding protein levels in British women.
Cancer Causes Control
2003
;
14
:
65
–74.
45
Frystyk J, Vestbo E, Skjaerbaek C, Mogensen CE, Orskov H. Free insulin-like growth factors in human obesity.
Metabolism
1995
;
44
:
37
–44.
46
Mohamed-Ali V, Pinkney JH, Panahloo A, Cwyfan-Hughes S, Holly JM, Yudkin JS. Insulin-like growth factor binding protein-1 in NIDDM: relationship with the insulin resistance syndrome.
Clin Endocrinol (Oxf)
1999
;
50
:
221
–8.
47
Frystyk J, Skjaerbaek C, Vestbo E, Fisker S, Orskov H. Circulating levels of free insulin-like growth factors in obese subjects: the impact of type-2 diabetes.
Diabetes Metab Res Rev
1999
;
15
:
314
–22.
48
Kaaks R, Bellati C, Venturelli E, et al. Effects of dietary intervention on IGF-I and IGF-binding proteins, and related alterations in sex steroid metabolism: the Diet and Androgens (DIANA) Randomised Trial.
Eur J Clin Nutr
2003
;
57
:
1079
–88.
49
Irwin ML, McTiernan A, Bernstein L, et al. Relationship of obesity and physical activity with C-peptide, leptin, and insulin-like growth factors in breast cancer survivors.
Cancer Epidemiol Biomarkers Prev
2005
;
14
:
2881
–8.
50
Morimoto LM, Newcomb PA, White E, Bigler J, Potter JD. Variation in plasma insulin-like growth factor-1 and insulin-like growth factor binding protein-3: personal and lifestyle factors (United States).
Cancer Causes Control
2005
;
16
:
917
–27.
51
Teramukai S, Rohan T, Eguchi H, Oka T, Shinchi K, Kono S. Behavioral correlates of insulin-like growth factor binding protein-3 in middle-aged Japanese men.
Am J Epidemiol
2002
;
156
:
344
–8.
52
Chang S, Wu X, Yu H, Spitz MR. Plasma concentrations of insulin-like growth factors among healthy men and postmenopausal women: associations with body composition, lifestyle, and reproductive factors.
Cancer Epidemiol Biomarkers Prev
2002
;
11
:
758
–66.
53
Lukanova A, Lundin E, Zeleniuch-Jacquotte A, et al. Body mass index, circulating levels of sex-steroid hormones, IGF-1, and IGF-binding protein-3: a cross-sectional study in healthy women.
Eur J Endocrinol
2004
;
150
:
161
–70.
54
Schoen RE, Schragin J, Weissfeld JL, et al. Lack of association between adipose tissue distribution and IGF-I and IGFBP-3 in men and women.
Cancer Epidemiol Biomarkers Prev
2002
;
11
:
581
–6.
55
Holmes MD, Pollak MN, Willett WC, Hankinson SE. Dietary correlates of plasma insulin-like growth factor I and insulin-like growth factor binding protein 3 concentrations.
Cancer Epidemiol Biomarkers Prev
2002
;
11
:
852
–61.