| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
1 Orentreich Foundation for the Advancement of Science, Inc., 855 Route 301, Cold Spring-on-Hudson, NY 10516
aauthor for correspondence: fax 845-265-4210, e-mail ofas1{at}juno.com
Recent studies have correlated high serum insulin-like growth factor-1 (IGF-1) with increased risk of several common cancers, primarily prostate, breast, colorectum, and lung (1)(2)(3); low concentrations of circulating IGF-1 have been associated with osteoporosis (4), impaired cognitive function (5), and heart disease (6). IGF-1 screening may help prevent chronic disease by identifying high-risk individuals (7)(8). Although normal adult IGF-1 concentrations relative to age and sex have been widely reported, based on cross-sectional data (9)(10)(11)(12)(13)(14), there have been few longitudinal studies. Most studies measuring multiple samples followed only a few individuals over short periods (15)(16)(17)(18)(19). Knowing the degree of individual variation over longer time periods would be useful (20).
Staff volunteers gave informed consent for the study. Blood was drawn monthly between 0800 and 1000 in the morning from 26 apparently healthy adults, and medications and time of last meal were recorded. Within 1 h of collection, samples were centrifuged, and serum was stored at -20 °C until analysis, which was within 1 month of collection. Serum IGF-1 was measured in 13 men (starting ages, 24–77 years; median, 49 years) and 13 women (starting ages, 23–56 years; median, 41 years). The total number of samples from each individual varied from 13 to 56 (mean, 31), and they were collected over 1.2–6 years (mean, 4.5 years). Use of previously collected serum stored at -40 °C extended the time period followed for two male volunteers to 16 years. In addition, multiple aliquots from three male volunteers (ages, 26, 44, and 75 years) used for assessing interassay variability were stored at -20 °C for 1–12 months.
IGF-1 concentrations were measured by IRMAs from Diagnostic Systems Laboratories. Samples were extracted with acid-ethanol and assayed in duplicate. The intraassay CV was 3.0%. Interassay variability over 6 years was 10%, calculated from the weighted mean CV of the male aliquots mentioned and control pools from the assay manufacturer and Bio-Rad Laboratories. Adult IGF-1 reference intervals were 9.8–91.5 nmol/L in men and women <51 years of age and 6.5–39.2 nmol/L for ages 
51 years; these reference intervals were established using averaged duplicate measurements in 250 patients over eight assays.
IGF-1 reproducibility over 1 and 5 years was evaluated using the intraclass correlation coefficient (ICC) (21). ICC (
) is the proportion of total variance of an observation associated with its class membership [i.e., = 
2b/(2b + 2w), where 2b is between-person variability and 2w is within-person variability]. Thus, ICC is used as a measure of reproducibility of replicate measures from the same individual. ICC values 0.75 represent excellent reproducibility; values between 0.40 and 0.75 represent fair to good reproducibility (22)(23). Category-specific ICCs were compared by the method described by Donner and Wells (24). To determine trends in IGF-1 measures over time, either overall or within age or gender groups, a repeated-measures ANOVA was used. Corresponding significance levels using F-tests are provided for group, time, and group-by-time interactions. P values associated with time effects and group-by-time interactions are reported with the Greenhouse–Geisser adjustment correcting for lack of sphericity. Statistical tests (using Stata v.7) achieving critical levels of <0.05 were considered significant.
Each participant’s IGF-1 value was plotted against age at the time of sample collection (Fig. 1A
) and revealed a negative correlation with age: y = -27.628ln(x) + 143.67; SE = 0.40. Two male volunteers, male 11 (age, 62 years; mean IGF-1, 24.8 ± 4.4 nmol/L) and male 12 (age, 64 years; mean IGF-1, 12.8 ± 2.1 nmol/L), were followed long enough to merit separate linear regression analyses on their data (Fig. 1B
). Their concentrations also showed a negative correlation with age over 16 years [male 11, y = -26.706ln(x) + 140.11 (SE = 0.58); male 12, y = -8.8335ln(x) + 50.934 (SE = 0.25)].
|
Mean age and IGF-1 concentrations along with CVs were calculated over all observations for each individual. The mean concentration for all participants (mean age, 47.5 years; range, 25.5–80.3 years) averaged from individual means was 38.0 ± 13.5 nmol/L. There was no appreciable difference between male and female mean values (38.0 ± 14.1 and 37.9 ± 12.7 nmol/L, respectively), although female participants were younger (41.3 vs 53.6 years). However, younger participants had a higher mean IGF-1 concentration (age <45 years, 48.0 ± 7.5 nmol/L; age 45 years, 29.5 ± 11.4 nmol/L; P = 0.0001). The average participant CV was 19% (range, 12–29%).
IGF-1 reproducibility in individuals was assessed across 1 and 5 years (Table 1
). The measurement variability at different intervals was removed by selecting only three equidistant time points within each period: baseline, mid-point, and end. All 26 participants were included in the 1-year assessment, and 16 individuals (10 men and 6 women) were included in the 5-year assessment. Separate assessments by both age and gender were also made.
|
IGF-1 ICCs for 1 and 5 years were 0.79 [95% confidence interval (CI), 0.66–0.91] and 0.74 (95% CI, 0.55–0.93), respectively. The ICC was calculated separately by gender for both time periods. Interestingly, the male ICC over 1 year was 0.90 (95% CI, 0.81–0.99), whereas among women it was appreciably lower at 0.69 (95% CI, 0.46–0.93). Although the difference was smaller between genders, the 5-year male ICC of 0.77 (95% CI, 0.55–0.99) was also higher (female, 0.71; 95% CI, 0.36–1.0). Neither 1- nor 5-year gender differences achieved statistical significance (P = 0.12 and 0.78, respectively).
There were no statistical differences by age category in 1- and 5-year ICCs. For 1 year, the age <45 ICC was 0.69 (95% CI, 0.46–0.92), whereas the age 45 ICC was 0.71 (95% CI, 0.47–0.94; P = 0.91). The 5-year ICC (age <45) was 0.64 (95% CI, 0.29–0.98) compared with 0.72 (95% CI, 0.43–1.00) for age 45 (P = 0.73).
One- and 5-year concentrations were also analyzed for age (<45 vs 45), gender, and time effects. Separate repeated-measures ANOVA showed neither a main gender effect at either 1 or 5 years (P = 0.79 and 0.81, respectively) nor evidence of consistent change in mean IGF-1 over time over either 1 year (P = 0.43) or 5 years (P = 0.36). There was a significant main effect for age (P <0.001 and P = 0.03, respectively) for both time periods. Individuals <45 years of age had higher concentrations than individuals 45 years (mean difference over 1 year, 15.6 nmol/L; mean difference over 5 years, 16.6 nmol/L). Finally, for trends within demographic characteristics, time-by-age group and time-by-gender interactions were examined. For 1 and 5 years, there was no evidence of time-by-gender (P = 0.34 and 0.23, respectively) or time-by-age group interactions (P = 0.79 and 0.65, respectively).
Researchers measuring more than one IGF-1 concentration in an individual have found excellent agreement between measurements, comparing either two values (2)(3)(15) or multiple values over short periods (16)(17)(18)(19). Our data show that during 1- and 5-year intervals, IGF-1 concentrations remain stable; however, individual IGF-1 variation, as expressed in the average participant’s CV of 19%, was greater than expected.
The data were analyzed to determine whether within-person variability affected the reliability of IGF-1 as a biomarker. Both the 1- and 5-year ICCs (0.79 and 0.74, respectively) confirmed that most of the IGF-1 variability was between individuals rather than within individuals, a critical aspect for ascertaining disease associations with serum IGF-1 concentrations. Although this study had only 26 participants, between-person variances as reflected by the mean and 2 SD range (age <45 years, 48 ± 15 nmol/L; age 45 years, 29.5 ± 22.8 nmol/L) fell comfortably within the reference interval established in our clinical laboratory and agree with the age-specific means reported by Yu et al. (10), indicating that the between-person variances reported here are reasonable representations. In addition, variability from analytical error in this study was small (10%), although IGF-1 concentrations were measured in different assays. The reliability coefficients reported here are in the range of 0.6–0.9 and are comparable to other biomarkers (25)(26).
Reliability assessments, calculated separately by age and gender, showed that age does not affect the reliability of IGF-1 measurements. However, although not statistically significant, our findings of a male ICC of 0.90 vs female ICC of 0.69 suggest a difference in the variability of IGF-1 between men and women. It is unlikely that the menstrual cycle influenced IGF-1 variability (17)(27), but oral estrogen replacement therapy is known to decrease IGF-1 concentrations (28)(29). Several women took oral estrogens during the study. However, the small number of participants and the study design are inadequate to address whether the greater female variability of IGF-1 seen in this study depends on gender, estrogen usage, or sampling variability.
The lower IGF-1 concentrations shown here in older adults agree with previous findings. Both the negative slope of the regression by age and the lower mean concentration agree with others (9)(10)(11)(12)(13)(14)(15). However, when concentrations in individuals were examined, IGF-1 concentrations were stable over both 1- and 5-year time periods. In two older men, IGF-1 concentrations over 16 years gradually decreased with age, although both mean concentrations and rates were dissimilar (male 11, 24.8 ± 4.4 nmol/L; male 12, 12.8 ± 2.1 nmol/L). Previous studies of multiple IGF-1 concentrations in individuals established that concentrations vary little during the day and are unaffected by prandial status (16)(17)(18). We have shown here that they also remain relatively stable over a period of years.
In summary, participants maintained relatively constant IGF-1 concentrations over time. There were discernable differences between individuals, and assay error was negligible. A single IGF-1 determination is reliable and can be used in estimating disease risk after age adjustment.
Acknowledgments
We thank Gregory Cornelius, Leslie Mauchline, and LuAnna Tardio for technical assistance and Angela Tremain, Nancy P. Durr, Steve Massardo, and Sylvia Duffy for support in manuscript preparation. We also acknowledge Dr. Paul Visintainer for assistance with the statistical analysis.
References
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  S. H. Siahpush, T. L. Vaughan, J. N. Lampe, R. Freeman, S. Lewis, R. D. Odze, P. L. Blount, K. Ayub, P. S. Rabinovitch, B. J. Reid, et al. Longitudinal Study of Insulin-like Growth Factor, Insulin-like Growth Factor Binding Protein-3, and their Polymorphisms: Risk of Neoplastic Progression in Barrett's Esophagus Cancer Epidemiol. Biomarkers Prev., November 1, 2007; 16(11): 2387 - 2395. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Baglietto, D. R. English, J. L. Hopper, H. A. Morris, W. D. Tilley, and G. G. Giles Circulating Insulin-Like Growth Factor-I and Binding Protein-3 and the Risk of Breast Cancer Cancer Epidemiol. Biomarkers Prev., April 1, 2007; 16(4): 763 - 768. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. dos Santos Silva, N. Johnson, B. De Stavola, G. Torres-Mejia, O. Fletcher, D. S. Allen, N. E. Allen, T. J. Key, I. S. Fentiman, J. M.P. Holly, et al. The insulin-like growth factor system and mammographic features in premenopausal and postmenopausal women. Cancer Epidemiol. Biomarkers Prev., March 1, 2006; 15(3): 449 - 455. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
, male 11; 
, male 12.