Skip Navigation


American Journal of Epidemiology Advance Access originally published online on June 29, 2007
American Journal of Epidemiology 2007 166(5):518-526; doi:10.1093/aje/kwm124
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
166/5/518    most recent
kwm124v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (8)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Saydah, S.
Right arrow Articles by Berrigan, D.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Saydah, S.
Right arrow Articles by Berrigan, D.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

American Journal of Epidemiology Published by the Johns Hopkins Bloomberg School of Public Health 2007.

ORIGINAL CONTRIBUTIONS

Insulin-like Growth Factors and Subsequent Risk of Mortality in the United States

Sharon Saydah1, Barry Graubard2, Rachel Ballard-Barbash3 and David Berrigan3

1 Office of Analysis and Epidemiology, National Center for Health Statistics, Centers for Disease Control and Prevention, Hyattsville, MD
2 Biostatistics Branch, Division of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, MD
3 Applied Research Program, Division of Cancer Control and Population Sciences, National Cancer Institute, Bethesda, MD

Correspondence to Dr. Sharon Saydah, Office of Analysis and Epidemiology, National Center for Health Statistics, Centers for Disease Control and Prevention, 3311 Toledo Road, Hyattsville, MD 20782 (e-mail: sharon{at}saydah.com; zle0{at}cdc.gov).

Received for publication October 30, 2006. Accepted for publication March 15, 2007.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
Although numerous studies have explored the relation of insulin-like growth factor (IGF)-I and IGF-binding protein (BP) 3 with cancer and cardiovascular disease, only two previous studies are known to have looked at the association of IGF-I and IGF-BP3 with risk of mortality. The objective of this US study was to examine the risk of all-cause, heart disease, and cancer mortality associated with IGF-I and IGF-BP3 levels using data from the Third National Health and Nutrition Examination Survey (NHANES III) and NHANES III Mortality Study (n = 6,061) (1988–2000). The authors constructed proportional hazards models with age as the time scale to determine the association of baseline IGF-I and IGF-BP3 levels with subsequent mortality. After adjustment for baseline measures, there was no increased risk of all-cause, heart disease, or cancer mortality for the lower quartiles of IGF-I compared with the highest quartile. The adjusted relative hazard of all-cause mortality for the lowest quartile of IGF-BP3 compared with the highest quartile was 1.57 (95% confidence interval: 0.98, 2.52), and the trend for risk was significant (p = 0.0364), but there was no increased risk of heart disease or cancer mortality. Results suggest that the association of IGF-I and IGF-BP3 with mortality may differ from associations with incidence of disease.

heart diseases; insulin-like growth factor I; insulin-like growth factor binding protein 3; mortality; neoplasms; nutrition surveys

Abbreviations: BP, binding protein; ICD-9, International Classification of Diseases, Ninth Revision; ICD-10, International Classification of Diseases, Tenth Revision; IGF, insulin-like growth factor; NHANES III, Third National Health and Nutrition Examination Survey


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
Insulin-like growth factors (IGFs) and their associated binding proteins (BPs) play important roles in normal development and growth (1). Furthermore, the IGF system has been implicated in the development of cardiovascular disease and cancer (1–3). Biologic effects and bioavailability of IGF-I are modulated through IGF-BPs, which control IGF-I access to cell surface receptors (1–3).

The relation of the IGF-I and IGF-BPs with human disease is complex. Most epidemiologic studies have focused on IGF-I, IGF-BP1, and IGF-BP3. Whereas abnormally high levels of IGF-I are associated with acromegaly (4, 5), the associations with disease within the normal range of IGF-I are less clear. Higher IGF-I has been associated with a moderately increased risk of cancer (6), particularly prostate, colorectal, and breast cancers. However, the direction of the relation of IGF-I with cardiovascular disease is opposite from the relation with incident cancers. Lower IGF-I levels are associated with atherosclerotic plaque (7–9), development of ischemic heart disease during 15 years of follow-up (7, 10), and myocardial infarction (7, 11–13).

Previous studies of IGF-I and IGF-BP3 and subsequent cancer mortality have focused on specific patient populations (14) and the use of IGF-I and/or IGF-BP3 for prediction of cancer survival, and prognosis (15–21). Few previous studies have examined the risk of mortality in the general population in relation to IGF-I levels, and no study, to our knowledge, has examined mortality in relation to IGF-BP3 levels. An increased risk of cardiovascular disease mortality was associated with lower IGF-I levels at baseline in the Rancho Bernardo prospective cohort study (22). In a cohort of Finnish men, higher levels of IGF-BP1 were associated with increased risk of death from all causes and from cardiovascular disease (23). To date, no nationally representative prospective cohorts are known to have examined the risk of all-cause mortality based on IGF-I and IGF-BP3 levels. We conducted this cohort study to determine the risk of all-cause, heart disease, and cancer mortality associated with IGF-I and IGF-BP3 levels in the general population.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 References
 
Data were derived from the Third National Health and Nutrition Examination Survey (NHANES III) Mortality Study, a prospective cohort study that passively followed participants aged 17 years or older in NHANES III (n = 20,024). NHANES III was conducted between 1988 and 1994 by the National Center for Health Statistics. A stratified multistage sample design was used to produce a nationally representative sample of the noninstitutionalized US civilian population (24). The survey included a physical examination, laboratory test, and questionnaires with health- and nutrition-related topics. The overall response rate for adults aged 20 years or older for completing the interview and examination was 77 percent (25). Of individuals aged 20 years or older, 6,226 were selected at random for the morning examination and had surplus serum samples available for analysis of IGF-I and IGF-BP3.

Baseline assessments
Participants' age, sex, race/ethnicity, and personal health characteristics were obtained by interview. Physical examination included measuring height, weight, waist circumference, and hip circumference. Body mass index was calculated as kilograms per square meter for each participant.

Measurement of IGFs
IGF-I and IGF-BP3 were measured in the stored serum from the NHANES III morning fasting sample using the DSL Laboratories (Webster, Texas) IGF-I ELISA (catalog #10-5600) and IGF-BP3 IRMA (catalog #6600). All assays described were performed by a single technician at the DSL facility in Webster, Texas. For the IGF-I ELISA, a single batch of reagents sufficient for the entire experiment was frozen at study onset. The IGF-BP3 IRMA required fresh batches of radioactive tracer during the study. Throughout the study, samples were reanalyzed if the coefficient of variation for replicate samples from a single vial was greater than 15 percent. Complete details on the laboratory methods used and quality control measures have been reported previously (26).

Outcomes
NHANES III participants aged 17 years or older for whom data were available for matching were matched to the National Death Index to determine mortality status. The National Death Index was searched through December 31, 2000, for follow-up. NHANES III and the National Death Index are linked by probabilistic matching. The National Center for Health Statistics conducted the linkage and created scores for potential matches. For a selected sample of NHANES III records, the Center reviewed the death certificate record to verify correct matches. Overall, 20,024 adult NHANES III participants were eligible for mortality follow-up by linkage with the National Death Index, of whom 3,384 were identified as deceased. A complete description of the methodology used to link NHANES III records to the National Death Index can be found at the following website: http://www.cdc.gov/nchs/data/datalinkage/matching_methodology_nhanes3_final.pdf (accessed April 25, 2006).

Underlying cause of death is based on International Classification of Diseases, Ninth Revision (ICD-9) codes from 1986 to 1998 and on International Classification of Diseases, Tenth Revision (ICD-10) codes from 1999 to 2000. Heart disease deaths were defined as those with ICD-9 codes 390–398, 402, and 404–429 and ICD-10 codes I00–I09, I11, I13, and I20–I51. Cancer deaths were defined as those with ICD-9 codes 140–208 and ICD-10 codes C00–C097. Cause-of-death codes for heart disease and cancer based on ICD-9 and ICD-10 were selected for high comparability between the two coding methods (27).

For NHANES III participants with valid IGF data (n = 6,061), 5,313 were assumed to be alive, 743 were assumed to be deceased, and five were missing information for the linkage. Among those assumed deceased, two participants were missing information on cause of death. Overall, there were 743 deaths from all causes, 251 deaths from heart disease, and 181 deaths from cancer during 51,238 person-years of follow-up. Person-years of follow-up were calculated for each participant based on the end of follow-up (date of death for those assumed deceased or December 31, 2000, for those assumed alive) minus the date of the NHANES III examination.

Analysis
All analyses were weighted to the US population to provide nationally representative estimates. SUDAAN statistical software, release 9.1 (28) was used to account for the complex survey design, a stratified multistage cluster sample (24, 25).

Participants' demographic characteristics, self-reported general health status, history of chronic conditions, body measurements, IGF-I, and IGF-BP3 were assessed at baseline. The association of mortality with IGF-I and IGF-BP3 was estimated by using quartiles of each. Cutpoints for the quartiles were determined separately for males and females since preliminary analysis showed males to have higher mean IGF-I and lower mean IGF-BP3 levels compared with females in almost every age group.

Person years
Mortality per 1,000 person-years was calculated for each IGF-I and IGF-BP3 quartile based on the weighted number of deaths and person-years. A log-linear Poisson model was used to calculate 95 percent confidence intervals.

Proportional hazards analysis
To determine whether differences in relative hazards between IGF-I and IGF-BP3 quartiles could be explained by other variables, proportional hazards models were constructed separately for IGF-I and IGF-BP3. Included were sex, race/ethnicity, smoking status, alcohol use, and body mass index.

The biologic interaction between IGF-I and IGF-BP3 is complex. The bioavailability of IGF-I and IGF-BP3 is measured by the amount of IGF-I or IGF-BP3 not bound to one another (2, 3, 5). Therefore, to account for the bioavailability of IGF-I and IGF-BP3, we included IGF-BP3 as a continuous variable in the final proportional hazards model for IGF-I and IGF-I as a continuous variable in the final proportional hazards model for IGF-BP3.

We used age as the timescale for analysis, with left truncation. Participants who entered at their age at examination were censored at the end of follow-up if they were still alive. For cause-specific analyses (i.e., cancer mortality or cardiovascular mortality), if a participant died from causes other than the specific cause of death of interest, then they were censored at the age at death. When participants with a baseline history of cancer (n = 401) were excluded from the analysis, there was no change in the direction or magnitude of the results. When participants with a baseline history of heart disease (n = 291) were excluded from the analysis, there was no change in the direction or magnitude of the results. Similarly, when participants who died within the first 2 years of follow-up were excluded from the analysis (n = 139), there was also no change in the direction or magnitude of the results. Therefore, the analysis included participants with a baseline history of cancer, heart disease, and death occurring within the first 2 years of follow-up. We found no significant first-order interactions between IGF quartile and any covariate (p > 0.05).

To check whether the functional relation of death and IGF-I was similar to the quartile analysis, we used the proportional hazards function to model IGF-I using a spline regression with three knots (29). We observed patterns similar to those when the quartile analysis was used. This analysis, checking for interactions and spline regression, was repeated with IGF-BP3. There were no significant first-order interactions between IGF-BP3 quartile and any covariate (p > 0.05). As for the quartile analysis, similar patterns were observed with the proportional hazards function to model IGF-BP3 using the spline regression.

The test for trend in the relation of IGF-I and IGF-BP3 with mortality was analyzed three ways: by entering the IGF-I and IGF-BP3 quartiles as ordinal variables in the proportional hazards models by assigning scores of 1, 2, 3, or 4 to the quartiles in the regression model; and by entering the median of each quartile in the proportional hazards models and as a continuous variable using a three-knot regression spline in the proportional hazards models. Similar results were obtained with each method. The p for trend from entering the quartiles as an ordinal variable is reported in the Results section below.

Because IGF-I and IGF-BP3 levels decline with age and the majority of deaths occurred at the older ages, we performed subsidiary analysis by restricting the cohort to participants aged 50 years or older at the time of NHANES III to determine the associated risk in a group of older adults.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 References
 
Baseline characteristics
Table 1 reports the baseline characteristics of participants included in our analysis overall and for males and females separately. Overall, mean age of the cohort at baseline was 44 years and was similar for males and females. Reported alcohol use differed by sex, with 24 percent of males reporting three or more drinks per week, but only 10 percent of females reporting the same. Similar differences by sex were found for smoking history, with 31 percent of males but only 24 percent of females reporting they were current smokers. As expected, waist circumference and waist-to-hip ratio differed by sex; females had lower mean waist circumferences and mean waist-to-hip ratios. IGF-I and IGF-BP3 also differed by sex. Compared with females, males had higher IGF-I levels and lower IGF-BP3 levels.


View this table:
[in this window]
[in a new window]

  		TABLE 1. Baseline characteristics of 6,056 adults aged 20 years or older in the Third National Health and Nutrition Examination Survey, United States, 1988–1994*

 
IGF-I and mortality
Table 2 presents the results by IGF-I quartile and the risk of death from all causes, heart disease, and cancer. Mortality decreased with increasing IGF-I quartiles for deaths from all causes, heart disease, and cancer. Mortality from all causes was highest for participants in the lowest IGF-I quartile, with a mortality rate per 1,000 person-years of 17.8 (95 percent confidence interval: 15.34, 20.24), decreasing in the second quartile to 10.2 (95 percent confidence interval: 7.97, 12.43), to 5.9 (95 percent confidence interval: 4.44, 7.34) in the third quartile, and to 2.8 (95 percent confidence interval: 2.02, 3.62) in the fourth quartile. This trend was statistically significant (p < 0.001). After we adjusted for age, the relative hazard point estimates and the trend were no longer statistically significant. Adjusting for other covariates (sex, race/ethnicity, smoking status, alcohol use, and body mass index) did not substantially change the results. Adjusting for IGF-BP3, to estimate the bioavailability of IGF-I, also did not change the results. The results from the proportional hazards model and a three-knot spline are displayed graphically in figure 1. The confidence intervals are wide, and the risk of all-cause mortality did not increase with increasing IGF-I levels.


View this table:
[in this window]
[in a new window]

  		TABLE 2. All-cause, heart disease, and cancer mortality by insulin-like growth factor-I quartiles for adults aged 20 years or older in the Third National Health and Nutrition Examination Survey Mortality Study, United States, 1988–2000

 

Figure 1
View larger version (8K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		FIGURE 1. Relative hazard of all-cause mortality for different insulin-like growth factor-I levels compared with the referent of 14.48 ng/ml–10 (the 12.5th percentile, as indicated by the vertical line) among adults aged 20 years or older in the United States, Third National Health and Nutrition Examination Survey Mortality Study, 1988–2000. The solid horizontal line shows the fitted three-knot spline relation; the dashed lines are the point-wise upper and lower 95% confidence limits.

 
Similar to the patterns observed for all-cause mortality, the highest rates for heart disease and cancer mortality were observed in the lowest quartile of IGF-I and decreased with increasing quartiles (p for trend < 0.001). After we adjusted for age or for age and the other covariates in proportional hazards models, the relative hazard for heart disease mortality was no longer statistically significant.

A slightly different pattern was observed with cancer mortality. Although the relative hazards for cancer mortality after adjustment were no longer statistically significant, the point estimates for the lower quartiles compared with the highest quartile suggested an increased risk of cancer mortality. The trend was not statistically significant for any of the adjusted proportional hazards models (p for trend > 0.05).

IGF-BP3 and mortality
Table 3 presents the results by IGF-BP3 quartile and the risk of death from all causes, heart disease, and cancer. Similar to patterns observed with IGF-I, mortality decreased with increasing IGF-BP3 quartile for deaths from all causes, heart disease, and cancer.


View this table:
[in this window]
[in a new window]

  		TABLE 3. All-cause, heart disease, and cancer mortality by insulin-like growth factor-binding protein 3 quartiles for adults aged 20 years or older in the Third National Health and Nutrition Examination Survey Mortality Study, United States, 1988–2000

 
Mortality from all causes ranged from 18.12 per 1,000 person-years (95 percent confidence interval: 15.02, 21.22) in the lowest IGF-BP3 quartile to 3.85 per 1,000 person-years (95 percent confidence interval: 2.79, 4.91) in the highest IGF-BP3 quartile. This trend was statistically significant. After adjustments, the trend for decreasing mortality from the lowest to the highest quartile remained significant. In model 3, adjusted for age, sex, race/ethnicity, smoking status, alcohol use, body mass index, and IGF-I, compared with that for the highest quartile, the relative hazard for the lowest IGF-BP3 quartile was 1.57 (95 percent confidence interval: 0.98, 2.52), followed by quartile 2 (relative hazard = 1.40, 95 percent confidence interval: 0.95, 2.05) and quartile 3 (relative hazard = 1.09, 95 percent confidence interval: 0.65, 1.81). The results from the proportional hazards model and a three-knot spline are displayed graphically in figure 2. The confidence intervals are wide, but the risk of all-cause mortality seemed to decrease with increasing IGF-BP3 levels.


Figure 2
View larger version (9K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		FIGURE 2. Relative hazard of all-cause mortality for different insulin-like growth factor binding protein-3 levels compared with the referent of 0.033 ng/ml–100,000 (the 12.5th percentile, as indicated by the vertical line) among adults aged 20 years or older in the United States, Third National Health and Nutrition Examination Survey Mortality Study, 1988–2000. The solid horizontal line shows the fitted three-knot spline relation; the dashed lines are the point-wise upper and lower 95% confidence limits.

 
A similar pattern was observed for heart disease and cancer mortality with IGF-BP3. For both heart disease and cancer mortality, in unadjusted analysis, increasing IGF-BP3 quartiles were associated with decreasing mortality. However, after multivariate adjustments, the relative hazard estimates were not statistically significant and the trends were also not significant.

Subsidiary analysis among adults aged 50 years or older
We repeated the analysis by restricting the cohort to participants aged 50 years or older at baseline. After we adjusted for age, sex, race/ethnicity, smoking status, alcohol use, and body mass index, there was no increased risk with decreasing IGF-I level for all-cause, heart disease, or cancer deaths. The results for IGF-BP3 and all-cause mortality were similar to those for the full cohort. After adjustments, the relative hazard point estimates for each quartile compared with quartile 4 were not significant for quartile 1 (relative hazard = 1.72, 95 percent confidence interval: 0.98, 3.01) or quartile 3 (relative hazard = 1.29, 95 percent confidence interval: 0.70, 2.39), but the estimate was significant for quartile 2 (relative hazard = 1.61, 95 percent confidence interval: 1.02, 2.55) and the test for trend was significant (p = 0.04). There were no significant associations with IGF-BP3 and heart disease or cancer deaths after adjustments.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
References
 
In this nationally representative sample of adults, lower levels of IGF-BP3 were associated with increased risk of all-cause mortality, but not heart disease mortality, after adjusting for age, sex, race/ethnicity, smoking status, alcohol use, body mass index, and IGF-I. To our knowledge, no previous studies have examined the relation of IGF-BP3 and all-cause or heart disease mortality in the general population. One previous study examined the relation of IGF-BP1, but not IGF-BP3, in a cohort of Finnish men (23). In the Finnish study, higher levels of IGF-BP1 were associated with increased risk of death from all causes and from cardiovascular disease (23). In contrast, our study found that lower levels of IGF-BP3 were associated with higher mortality. Even though IGF-BP1 and IGF-BP3 are both BPs in the IGF axis, it appears that they may have different functions. IGF-BP1 correlates inversely with serum insulin (30, 31) and is strongly correlated with factors leading to atherogenesis (32). Unlike IGF-BP1, IGF-BP3 is primarily regulated by growth hormone. IGF-BP3 can directly promote or inhibit growth in vitro and in vivo and is the main BP for IGF-I (33). We did not measure IGF-BP1 in our study.

In our study, we did not find any significant associations between IGF-I and all-cause or heart disease mortality. In contrast to our study, a prospective study from the Rancho Bernardo cohort (22) found that lower levels of IGF-I were associated with an increased risk of cardiovascular disease mortality. Lower levels of IGF-I appear to be associated with an atherosclerotic process leading to cardiovascular disease (10). Previous cohort studies have shown that lower levels of IGF-I are associated with increased risk of incident cardiovascular disease events, which correlate highly with increased mortality (7–13).

We found no significant associations between IGF-I or IGF-BP3 and cancer mortality in this study. As far as we know, no previous studies have examined these associations with cancer mortality in the general population. IGF-I is thought to be associated with incident cancer because it stimulates growth and replication of carcinogenetic cells. However, different biologic mechanisms might operate for incident cancer and cancer mortality. Once cancer is diagnosed and treatment initiated, lower IGF-I levels may predict higher cancer mortality for this same reason—low levels of IGF-I make it more difficult for the body to stimulate growth and replication of healthy cells. Of note, only a few cancers have been associated with higher IGF-I levels in epidemiologic studies (2, 3, 6), but no study has looked specifically at cancer mortality. For example, higher IGF-I levels were associated with colorectal cancer but not lung cancer in a recent meta-analysis of previously published studies (6). In this study, lack of association with cancer mortality may also be because all cancers were considered together.

The associations of IGF-I and IGF-BP3 with mortality are clearly confounded by age. In unadjusted analysis, lower levels of IGF-I and IGF-BP3 are associated with increased risk of death from all causes, heart disease, and cancer. However, once we adjusted for age, the association decreased toward the null and was often no longer significant. Increasing age is related to both decreasing IGF-I and IGF-BP3 levels and death. As adults age, levels of IGF-I and IGF-BP3 decline, with the lowest levels found in the oldest adults. Thus, adjusting for age in the proportional hazards analysis explained much of the association of decreasing IGF-I and IGF-BP3 with increased risk of death.

One possible reason why age has such a strong effect is the association of IGF levels with muscle mass. Decrease in IGF levels may correspond to an inability to maintain muscle mass during aging (34). Although loss of muscle mass is associated with aging, loss of muscle mass itself was not associated with mortality in a cohort of adults older than age 70 years in the United States (35). A recent review by Yang et al. (36) suggested that lower levels of IGF-I in early adulthood and higher levels of IGF-I later in life may be most beneficial for longevity. Further clouding the picture regarding the relation of IGF and mortality is that lower IGF-I levels have been shown to increase life expectancy (3, 37, 38). The relation of IGF-I and IGF-BP3 with age may also be influenced by other age-related factors such as sex hormones, body composition, and lifestyle (2). We attempted to control for some of these factors in the adjusted analysis but likely were unable to fully adjust for all potential age-related factors.

Strengths of this study include a nationally representative sample, 4–12 years of follow-up, and careful measurement of IGF-I and IGF-BP3 using current state-of-the-art methods. Nonetheless, there are two main limitations of this study. First, although follow-up of the cohort of participants was almost complete, there were few deaths, with fewer than 200 deaths from cancer. This limitation prevented us from looking at specific cancers or more refined death categories. Second, levels of IGF-I and IGF-BP3 were measured at only one point in time, at baseline. Single-point-in-time measurements assume that those for individuals with high levels of IGF-I or IGF-BP3 at baseline are consistently high and, conversely, that those for individuals with lower levels of IGF-I or IGF-BP3 are consistently low. IGF-I and IGF-BP3 levels may also vary day to day and over the longer term. This variation is expected to be minimal because two measurements of IGF-I from the same individual within 40 days of each other are highly correlated (39). Levels may vary based on whether or not an individual is fasting. Since IGF-I and IGF-BP3 were measured from serum samples collected after the participants fasted overnight, this variation is expected to be minimal.

In conclusion, in this nationally representative prospective study, we found no association of IGF-I with mortality from all causes, heart disease, or cancer. We did find an association of lower IGF-BP3 with increased all-cause mortality, but not heart disease or cancer mortality. The relation of IGF-I and IGF-BP3 with mortality is clearly confounded by age, and the biologic relation between each of them and human disease is complex. These results suggest that the association of IGF with mortality may differ from the previously observed associations with incidence of disease, particularly for cancer. The results also suggest that large samples or long durations of follow-up are necessary to accumulate sufficient outcomes to identify associations.


   ACKNOWLEDGMENTS
 
The views and interpretations presented in this paper are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention or the National Cancer Institute.

Conflict of interest: none declared.


   References
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 

  1. LeRoith D, Clemmons D, Nissley P, et al. NIH conference. Insulin-like growth factors in health and disease. Ann Intern Med (1992) 116:854–62.[Abstract/Free Full Text]
  2. Juul A. Serum levels of insulin-like growth factor I and its binding proteins in health and disease. Growth Horm IGF Res (2003) 13:113–70.[CrossRef][Web of Science][Medline]
  3. Pollak MN, Schernhammer ES, Hankinson SE. Insulin-like growth factors and neoplasia. Nat Rev Cancer (2004) 4:505–18.[CrossRef][Web of Science][Medline]
  4. Beers MH, Berkow R, eds. The Merck manual of diagnosis and therapy. (1999) 17th ed. Whitehouse Station, NJ: Merck Research Laboratories.
  5. Rosen CJ. Serum insulin-like growth factors and insulin-like growth factor-binding proteins: clinical implications. Clin Chem (1999) 45(8 pt 2):1384–90.[Abstract/Free Full Text]
  6. Renehan AG, Zwahlen M, Minder C, et al. Insulin-like growth factor (IGF)-I, IGF binding protein-3, and cancer risk: systematic review and meta-regression analysis. Lancet (2004) 363:1346–53.[CrossRef][Web of Science][Medline]
  7. Goodman-Gruen D, Barrett-Connor E, Rosen C. IGF-1 and ischemic heart disease in older people. J Am Geriatr Soc (2000) 48:860–1.[Web of Science][Medline]
  8. Janssen JA, Stolk RP, Pols HA, et al. Serum free IGF-I, total IGF-I, IGF-BP1 and IGF-BP3 levels in an elderly population: relation to age and sex steroid levels. Clin Endocrinol (Oxf) (1998) 48:471–8.[CrossRef][Medline]
  9. Janssen JA, Stolk RP, Pols HA, et al. Serum total IGF-I, free IGF-I, and IGFB-1 levels in an elderly population: relation to cardiovascular risk factors and disease. Arterioscler Thromb Vasc Biol (1998) 18:277–82.[Abstract/Free Full Text]
  10. Juul A, Scheike T, Davidsen M, et al. Low serum insulin-like growth factor I is associated with increased risk of ischemic heart disease: a population-based case-control study. Circulation (2002) 106:939–44.[Abstract/Free Full Text]
  11. Conti E, Andreotti F, Sciahbasi A, et al. Markedly reduced insulin-like growth factor-1 in the acute phase of myocardial infarction. J Am Coll Cardiol (2001) 38:26–32.[Abstract/Free Full Text]
  12. Reeves I, Abribat T, Laramee P, et al. Age-related serum levels of insulin-like growth factor-I, -II and IGF-binding protein-3 following myocardial infarction. Growth Horm IGF Res (2000) 10:78–84.[CrossRef][Web of Science][Medline]
  13. Vasan RS, Sullivan LM, D'Agostino RB, et al. Serum insulin-like growth factor I and risk for heart failure in elderly individuals without a previous myocardial infarction: the Framingham Heart Study. Ann Intern Med (2003) 139:642–8.[Abstract/Free Full Text]
  14. Denti L, Annoni V, Cattadori E, et al. Insulin-like growth factor 1 as a predictor of ischemic stroke outcome in the elderly. Am J Med (2004) 117:312–17.[CrossRef][Web of Science][Medline]
  15. Chan JM, Stampfer MJ, Ma J, et al. Insulin-like growth factor-I (IGF-I) and IGF binding protein-3 as predictors of advanced-stage prostate cancer. J Natl Cancer Inst (2002) 94:1099–106.[Abstract/Free Full Text]
  16. Lin Y, Tamakoshi A, Kikuchi S, et al. Serum insulin-like growth factor-I, insulin-like growth factor binding protein-3, and the risk of pancreatic cancer death. Int J Cancer (2004) 110:584–8.[CrossRef][Web of Science][Medline]
  17. Parker AS, Cheville JC, Janney CA, et al. High expression levels of insulin-like growth factor-I receptor predict poor survival among women with clear-cell renal cell carcinomas. Hum Pathol (2002) 33:801–5.[CrossRef][Web of Science][Medline]
  18. Petridou E, Skalkidou A, Dessypris N, et al. Insulin-like growth factor binding protein-3 predicts survival from acute childhood leukemia. Oncology (2001) 60:252–7.[CrossRef][Web of Science][Medline]
  19. Rocha RL, Hilsenbeck SG, Jackson JG, et al. Insulin-like growth factor binding protein-3 and insulin receptor substrate-1 in breast cancer: correlation with clinical parameters and disease-free survival. Clin Cancer Res (1997) 3:103–9.[Abstract]
  20. Wakai K, Ito Y, Suzuki K, et al. Serum insulin-like growth factors, insulin-like growth factor-binding protein-3, and risk of lung cancer death: a case-control study nested in the Japan Collaborative Cohort (JACC) Study. Jpn J Cancer Res (2002) 93:1279–86.[CrossRef][Web of Science]
  21. Yu H, Levesque MA, Khosravi MJ, et al. Insulin-like growth factor-binding protein-3 and breast cancer survival. Int J Cancer (1998) 79:624–8.[CrossRef][Web of Science][Medline]
  22. Laughlin GA, Barrett-Connor E, Criqui MH, et al. The prospective association of serum insulin-like growth factor I (IGF-I) and IGF-binding protein-1 levels with all cause and cardiovascular disease mortality in older adults: the Rancho Bernardo Study. J Clin Endocrinol Metab (2004) 89:114–20.[Abstract/Free Full Text]
  23. Harrela M, Qiao Q, Koistinen R, et al. High serum insulin-like growth factor binding protein-1 is associated with increased cardiovascular mortality in elderly men. Horm Metab Res (2002) 34:144–9.[Web of Science][Medline]
  24. Plan and operation of the Third National Health and Nutrition Examination Survey. (1988–1994) Series 1: programs and collection procedures. Vital Health Stat 1 1994;Jul:1–407.
  25. National Center for Health Statistics. Analytic and reporting guidelines: The Third National Health and Nutrition Examination Survey, NHANES III, 1988–94. Vol 32. (1996) Hyattsville, MD: National Center for Health Statistics, Centers for Disease Control and Prevention.
  26. Berrigan D, Potischman N, Dodd K, et al. Serum levels of insulin-like growth factor-1 and insulin-like growth factor-1 binding protein-3: quality control for studies of stored serum. Cancer Epidemiol Biomarkers Prev (2007) 16:1017–22.[Abstract/Free Full Text]
  27. Anderson RN, Minino AM, Hoyert DL, et al. Comparability of cause of death between ICD-9 and ICD-10: preliminary estimates. Natl Vital Stat Rep (2001) 49:1–32.[Medline]
  28. Research Triangle Institute. SUDAAN, release 9.1. (2004) Research Triangle Park, NC: Research Triangle Institute.
  29. Korn EL, Graubard BI. Analysis of health surveys. Wiley series in probability and statistics. (1999) New York, NY: John Wiley & Sons.
  30. Mohamed-Ali V, Pinkney JH, Panahloo A, et al. Insulin-like growth factor binding protein-1 in NIDDM: relationship with the insulin resistance syndrome. Clin Endocrinol (Oxf) (1999) 50:221–8.[CrossRef][Medline]
  31. Suikkari AM, Koivisto VA, Rutanen EM, et al. Insulin regulates the serum levels of low molecular weight insulin-like growth factor-binding protein. J Clin Endocrinol Metab (1988) 66:266–72.[Abstract/Free Full Text]
  32. Gibson JM, Westwood M, Young RJ, et al. Reduced insulin-like growth factor binding protein-1 (IGF-BP1) levels correlate with increased cardiovascular risk in non-insulin dependent diabetes mellitus (NIDDM). J Clin Endocrinol Metab (1996) 81:860–3.[Abstract]
  33. Ferry RJ Jr, Cerri RW, Cohen P. Insulin-like growth factor binding proteins: new proteins, new functions. Horm Res (1999) 51:53–67.[CrossRef][Web of Science][Medline]
  34. Goldspink G. Impairment of IGF-I gene splicing and MGF expression associated with muscle wasting. Int J Biochem Cell Biol (2006) 38:481–9.[CrossRef][Medline]
  35. Newman AB, Kupelian V, Visser M, et al. Strength, but not muscle mass, is associated with mortality in the health, aging and body composition study cohort. J Gerontol A Biol Sci Med Sci (2006) 61:72–7.[Abstract/Free Full Text]
  36. Yang J, Anzo M, Cohen P. Control of aging and longevity by IGF-I signaling. Exp Gerontol (2005) 40:867–72.[CrossRef][Web of Science][Medline]
  37. Berrigan D, Perkins SN, Haines DC, et al. Adult-onset calorie restriction and fasting delay spontaneous tumorigenesis in p53-deficient mice. Carcinogenesis (2002) 23:817–22.[Abstract/Free Full Text]
  38. Hursting SD, Lavigne JA, Berrigan D, et al. Calorie restriction, aging, and cancer prevention: mechanisms of action and applicability to humans. Annu Rev Med (2003) 54:131–52.[CrossRef][Medline]
  39. Goodman-Gruen D, Barrett-Connor E. Epidemiology of insulin-like growth factor-I in elderly men and women. The Rancho Bernardo Study. Am J Epidemiol (1997) 145:970–6.

Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 Eur J EndocrinolHome page
B. B Yeap, S A P. Chubb, K. K Y Ho, J. W S Setoh, K. A McCaul, P. E Norman, K. Jamrozik, and L. Flicker
IGF1 and its binding proteins 3 and 1 are differentially associated with metabolic syndrome in older men
Eur. J. Endocrinol., February 1, 2010; 162(2): 249 - 257.
[Abstract] [Full Text] [PDF]

Home page
 DiabetesHome page
M. P. Brugts, C. M. van Duijn, L. J. Hofland, J. C. Witteman, S. W.J. Lamberts, and J. A.M.J.L. Janssen
IGF-I Bioactivity in an Elderly Population: Relation to Insulin Sensitivity, Insulin Levels, and the Metabolic Syndrome
Diabetes, February 1, 2010; 59(2): 505 - 508.
[Abstract] [Full Text] [PDF]

Home page
 Endocr. Rev.Home page
S. Jogie-Brahim, D. Feldman, and Y. Oh
Unraveling Insulin-Like Growth Factor Binding Protein-3 Actions in Human Disease
Endocr. Rev., August 1, 2009; 30(5): 417 - 437.
[Abstract] [Full Text] [PDF]

Home page
 J. Clin. Endocrinol. Metab.Home page
S. Glaser, N. Friedrich, R. Ewert, C. Schaper, M. Nauck, M. Dorr, H. Volzke, S. B. Felix, A. Krebs, H. Wallaschofski, et al.
Association between Serum Insulin-Like Growth Factor (IGF) I and IGF Binding Protein-3 and Lung Function
J. Clin. Endocrinol. Metab., July 1, 2009; 94(7): 2452 - 2458.
[Abstract] [Full Text] [PDF]

Home page
 J. Clin. Endocrinol. Metab.Home page
N. Friedrich, R. Haring, M. Nauck, J. Ludemann, D. Rosskopf, E. Spilcke-Liss, S. B. Felix, M. Dorr, G. Brabant, H. Volzke, et al.
Mortality and Serum Insulin-Like Growth Factor (IGF)-I and IGF Binding Protein 3 Concentrations
J. Clin. Endocrinol. Metab., May 1, 2009; 94(5): 1732 - 1739.
[Abstract] [Full Text] [PDF]

Home page
 Eur Heart JHome page
J. Sierra-Johnson, R. M. Fisher, A. Romero-Corral, V. K. Somers, F. Lopez-Jimenez, J. Ohrvik, G. Walldius, M.-L. Hellenius, and A. Hamsten
Concentration of apolipoprotein B is comparable with the apolipoprotein B/apolipoprotein A-I ratio and better than routine clinical lipid measurements in predicting coronary heart disease mortality: findings from a multi-ethnic US population
Eur. Heart J., March 2, 2009; 30(6): 710 - 717.
[Abstract] [Full Text] [PDF]

Home page
 Eur J EndocrinolHome page
M. Andreassen, I. Raymond, C. Kistorp, P. Hildebrandt, J. Faber, and L. O. Kristensen
IGF1 as predictor of all cause mortality and cardiovascular disease in an elderly population
Eur. J. Endocrinol., January 1, 2009; 160(1): 25 - 31.
[Abstract] [Full Text] [PDF]

Home page
 J. Clin. Endocrinol. Metab.Home page
K. Casazza, P. B. Higgins, J. R. Fernandez, M. I. Goran, and B. A. Gower
Longitudinal Analysis of the Insulin-Like Growth Factor System in African-American and European American Children and Adolescents
J. Clin. Endocrinol. Metab., December 1, 2008; 93(12): 4917 - 4923.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
166/5/518    most recent
kwm124v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (8)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Saydah, S.
Right arrow Articles by Berrigan, D.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Saydah, S.
Right arrow Articles by Berrigan, D.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?