Abstract

Laboratory studies suggest that insulin-like growth factor I (IGF-I) promotes prostatic growth. The authors evaluated the association between benign prostatic hyperplasia and IGF-I and its binding protein IGFBP-3 in community-dwelling men to determine whether this laboratory finding is manifest at the population level. Participants (n = 471) were Olmsted County, Minnesota, Caucasian males aged 40–79 years in 1990. Urologic measures were assessed from the International Prostate Symptom Score, peak urinary flow rates, prostate volume, and serum prostate-specific antigen (PSA), and serum IGF-I and IGFBP-3 levels were measured. After adjustment for age, the relative odds (odds ratios) of an abnormal urologic measure in men with high versus low serum IGF-I levels were 0.98 (95% confidence interval (CI): 0.66, 1.45) for a symptom score of >7, 1.14 (95% CI: 0.72, 1.80) for a peak urinary flow rate of <12 ml/second, 1.11 (95% CI: 0.72, 1.72) for a prostate volume of >30 ml, and 0.71 (95% CI: 0.46, 1.09) for a PSA level of >1.4 ng/ml. A low IGFBP-3 level was associated with an enlarged prostate (odds ratio = 1.72, 95% CI: 1.05, 2.82), after simultaneous adjustment for IGF-I and age, but not with other urologic measures. These data do not provide evidence for an association between benign prostatic hyperplasia and serum IGF-I.

Received for publication September 11, 2002; accepted for publication November 14, 2002.

Benign prostatic hyperplasia is an important cause of morbidity in older men, yet little is known about its causes. The few established risk factors for the condition are increasing age and an intact androgen system. To our knowledge, despite this latter association, consistent positive associations between circulating androgen levels and urologic measures of benign prostatic hyperplasia have not been observed. This lack of an association may, in part, be due to the effects of intermediary factors that mediate androgen action and/or influence prostatic growth; included might be growth factors, for example, insulin-like growth factor (IGF) I and II and epidermal growth factor (1, 2).

IGF-I and IGF-II are important in the regulation, development, and proliferation of prostatic growth. Both IGF-I and IGF-II are produced in many tissues including prostatic cells (35) and are transported in the circulation bound to their carrier proteins, IGF binding proteins (IGFBPs). IGFs are potent mitogenic growth factors, exhibit antiapoptotic effects (6, 7), and may affect intraprostatic androgens (5). Serum IGF-I levels have been positively associated with prostate cancer (812) and also with increased benign prostatic proliferation (13, 14). However, it has been suggested that the reported positive association between IGF-I and benign prostatic hyperplasia could be due to ascertainment bias (15, 16); high IGF-I levels may increase benign prostatic growth, increase care-seeking for urologic symptoms, and thereby increase the likelihood that subclinical prostate cancer will be detected.

IGFBP-3, the most abundant of the IGFBPs, has also been associated with prostatic growth. Seventy-five percent of IGF-I is bound to IGFBP-3, 20–25 percent is bound to the other binding proteins (IGFBP 1, 2, 4, and 5), and less than 1 percent is carried in the unbound state in the circulation (17). Therefore, relative concentrations of IGFBP-3 may influence serum and, possibly, prostatic tissue levels of IGF-I. Importantly, although the IGFBP-3–IGF-I complex is a high molecular weight protein that cannot diffuse into tissues, complexes of IGF-I and other IGFBPs have a lower molecular weight and can traverse the capillary membrane into tissues (17, 18). Consequently, a decrease in IGFBP-3 levels may lead to greater binding of IGF-I to other IGFBPs, increased diffusion into tissues, and an increase in tissue IGF-I levels, resulting in increased prostatic growth. High levels of IGFBP-3 have been associated with a decreased risk of prostate cancer (8), but other studies have found no significant association between IGFBP-3 and prostate cancer risk (9, 12, 19). Higher levels of IGFBP-3 have also been associated with greater prostatic volume in African-American men (20).

While these previous studies provide some support for a potential role of IGF-I and IGFBP-3 in the development of benign prostatic hyperplasia, to our knowledge this potential association has not been fully evaluated in community-dwelling men. Since prostatic tissue IGF-I levels are influenced by serum IGF-I and IGFBP-3 levels (21), we hypothesized that high serum IGF-I and/or low IGFBP-3 levels should be associated with a greater likelihood of benign prostatic hyperplasia. This study provided the opportunity to test this hypothesis by using information on serum IGF-I and IGFBP-3 levels and surrogate measures of benign prostatic hyperplasia from a cohort of community-dwelling men.

MATERIALS AND METHODS

Study subjects

Subjects were participants in a longitudinal study of lower urinary tract symptoms and benign prostatic hyperplasia. The study selection details have been described previously (22, 23). A sampling frame of Caucasian males living in Olmsted County, Minnesota, and aged 40–79 years on January 1, 1990, was created by using electronic databases (24, 25). From this sampling frame, a random sample of men was selected, and their medical records were examined for exclusion criteria that included a history of prostatectomy, prostate cancer, or other known medical conditions, other than benign prostatic hyperplasia, that could impair normal voiding function. Of the 3,874 men eligible, 2,115 agreed to participate (55 percent participation rate). A 25 percent (n = 537) random subsample (clinic cohort) was selected from the 2,115 men for a more detailed clinical investigation, and 471 men participated (88 percent participation rate) (26).

Follow-up was performed biennially by using the same protocol as at baseline: questionnaires, determination of peak urinary flow rates (Qmax), and clinical measurements including transrectal ultrasound to determine prostate volume as well as a blood draw for serum measurements. For the first and second evaluations, men who refused to participate, had died, or were lost to follow-up were replaced by similarly aged men from the community (27) by using the same selection criteria as at baseline. The present study used serum measurements and assessments from the 1996 follow-up round to assess the cross-sectional relation between total serum IGF-I and IGFBP-3 and the urologic measures.

Clinic cohort measurements

Questionnaire and uroflow evaluation

At baseline, study subjects completed a previously validated questionnaire that included items similar to the International Prostate Symptom Score, and a composite score was estimated for each subject (28). Qmax was also measured in private by using a standard uroflowmeter, with subjects in a standing position.

Clinical evaluation

Each subject underwent a blood draw and determination of serum prostate-specific antigen (PSA) prior to any prostatic manipulation, including a digital rectal examination. A transrectal ultrasound was performed by using a 7.5-MHz, biplanar endorectal transducer. The anterior posterior, transverse, and superior inferior dimensions were measured and prostate volume was calculated by using the formula for a prolate ellipsoid: volume = 0.52 (transverse × anterior posterior × superior inferior) (29). Men found to have prostate cancer on the basis of this evaluation with biopsy confirmation were excluded from our analyses.

Laboratory measurements

Total serum IGF-I and IGFBP-3 levels were measured by using immunometric assay kits (Diagnostic System Laboratories, Inc., Webster, Texas). The interassay coefficients of variation were 6.9 percent at 5.8 ng/ml and 14.4 percent at 20.9 ng/ml for IGFBP-3 and 9 percent at 64 ng/ml and 6.2 percent at 157 ng/ml for IGF-I. Serum PSA levels were measured by using the Hybritech Tandem R assay (Hybritech Inc., San Diego, California), a solid, phase-two-site immunoradiometric assay. The coefficients of variation for measuring serum PSA in the Mayo Clinic laboratory (Rochester, Minnesota) are 6.22 percent at 2.88 ng/ml and 6.82 percent at 41.6 ng/ml.

Statistical analysis

Descriptive statistics were estimated for the 1996 measurements of serum IGF-I and IGFBP-3 (exposure variables) and for urologic measures of benign prostatic hyperplasia determined from the 1996 evaluation: International Prostate Symptom Score, Qmax, prostate volume, and serum PSA level (a surrogate measure of prostate size). Urologic measures were dichotomized on the basis of standard cutpoints (26, 28, 30) as a symptom score of ≤7 and >7 (moderate or severe symptoms), Qmax of ≥12 and <12 ml/second (low Qmax), and prostate volume of ≤30 ml and >30 ml (enlarged prostate); serum PSA was dichotomized at the median (≤1.4 and >1.4 ng/ml). Cutpoints for IGF-I and IGFBP-3 based on medians and tertiles were used to divide the study cohort into two and three groups, respectively. The associations between urologic measures and exposure variables were assessed by using Wilcoxon rank sum and Kruskall-Wallis tests of association, and odds ratios and 95 percent confidence intervals were estimated from logistic regression models. A test for trend across tertiles was examined by using logistic regression models. Since age could be a correlate of the serum markers, in which case adjusting for age could be inappropriate, the results with and without adjustment for age are presented in this paper. Analyses stratified by 10-year age groups were performed to assess potential effect modification by age, and homogeneity of the odds ratio was assessed by using the Breslow-Day test.

RESULTS

The distributions of and correlations between IGF-I and IGFBP-3 levels and urologic measures are presented in table 1. Among the 471 study subjects, mean values were 59.2 years for age, 8.0 for International Prostate Symptom Score, 19.4 ml/second for Qmax, 29.9 ml for prostate volume, and 1.4 ng/ml for serum PSA level. Significant negative correlations were found between serum IGFBP-3 level and prostate volume, serum PSA, International Prostate Symptom Score, and age, and a positive correlation was found between serum IGFBP-3 level and Qmax. There were also significant negative correlations between IGF-I and serum PSA levels and between IGF-I level and age, and a strong positive correlation was found between IGF-I and IGFBP-3 levels (rs= 0.61, p = 0.001). After adjustment for age, the correlations between IGF-I, IGFBP-3, and urologic measures were not statistically significant.

Benign prostatic hyperplasia surrogate measures and IGF-I

Mean serum IGF-I levels were significantly higher in men whose serum PSA was ≤1.4 ng/ml compared with >1.4 ng/ml (140.2 vs. 120.7, p = 0.002) (table 2). No significant associations were found between serum IGF-I level and symptom severity (p = 0.07), flow rates (p = 0.5), or prostate size (p = 0.26) (table 2). These results were corroborated in contingency table analyses. In bivariate analyses, a high serum IGF-I (IGF-I >median) level was negatively associated with a serum PSA level of >1.4 ng/ml (odds ratio (OR) = 0.58, 95 percent confidence interval (CI): 0.39, 0.86) (table 3). There were also negative but nonsignificant associations with symptom severity (OR = 0.81, 95 percent CI: 0.55, 1.17), Qmax (OR = 0.89, 95 percent CI: 0.58, 1.37), and prostate size (OR = 0.88, 95 percent CI: 0.59, 1.32) (table 3). With adjustment for the effects of age, the association between serum IGF-I and serum PSA levels was no longer statistically significant (table 3). No trends were observed with IGF-I level.

Benign prostatic hyperplasia surrogate measures and IGFBP-3

Mean serum IGFBP-3 levels were significantly lower in men with greater symptom severity (International Prostate Symptom Score >7), depressed Qmax (<12 ml/second), an enlarged prostate (>30 ml), or an elevated serum PSA level (>1.4 ng/ml) compared with men with a symptom score of ≤7, Qmax of >12 ml/second, prostate volume of <30 ml, or PSA level of <1.4 ng/ml (table 2). In addition, mean serum IGFBP-3 levels were lower across successively older groups of men. These results were corroborated in logistic regression analyses. In bivariate analyses, a low IGFBP-3 (<median) level was associated with an elevated relative odds of an enlarged prostate (OR = 1.90, 95 percent CI: 1.27, 2.89) and a PSA level of >1.4 ng/ml (OR = 1.71, 95 percent CI: 1.14, 2.55) (table 4). The point estimates for an International Prostate Symptom Score of >7 (OR = 1.22, 95 percent CI: 0.84, 1.78) and a low Qmax (OR = 1.48, 95 percent CI: 0.96, 2.28) were also elevated, but the confidence intervals included 1. For all outcomes, the point estimates decreased with adjustment for age and the associations with prostate volume and serum PSA were no longer statistically significant (table 4). Cutpoints based on tertiles of serum IGFBP-3 suggested a dose-response association with prostate volume (p for trend = 0.02): odds ratios for men in the lowest and the middle tertiles were 2.11 (95 percent CI: 1.28, 3.49) and 1.61 (95 percent CI: 0.97, 2.67), respectively, compared with men whose IGFBP-3 levels were in the highest tertile. With adjustment for age, these estimates decreased toward the null value to 1.32 (95 percent CI: 0.76, 2,30) and 1.07 (95 percent CI: 0.62, 1.85), respectively (p for trend = 0.31).

In multivariable models with simultaneous adjustment for serum IGFBP-3 and IGF-I, serum IGFBP-3 was significantly associated with prostate volume, but no other significant associations were found between serum IGF-I or IGFBP-3 levels and the other urologic measures. A low serum IGFBP-3 level was associated with an increased likelihood of an enlarged prostate (OR = 2.14, 95 percent CI: 1.34, 3.40) (table 5); there was no effect modification by age (p = 0.53). Point estimates were also elevated for the association of low serum IGFBP-3 level with low Qmax (OR = 1.57, 95 percent CI: 0.95, 2.59) and a serum PSA level of >1.4 ng/ml (OR = 1.43, 95 percent CI: 0.91, 2.25), but these associations were not statistically significant. Nonsignificant associations were observed between a higher IGF-I (>median) level and moderate or severe symptoms (OR = 0.86, 95 percent CI: 0.56, 1.32), a low Qmax (OR = 1.12, 95 percent CI: 0.68, 1.84), an enlarged prostate (OR = 1.27, 95 percent CI: 0.80, 2.02), and higher PSA levels (OR = 0.69, 95 percent CI: 0.44, 1.08). With adjustment for age, the association between serum IGFBP-3 level and prostate volume decreased, but it remained statistically significant (OR = 1.72, 95 percent CI: 1.05, 2.82).

Since IGFBP-3 determines, to some extent, bioavailable IGF-I, we stratified our analysis by serum IGFBP-3 level. For men whose serum IGFBP-3 levels were lower, the odds ratios for the association between serum IGF-I level and symptom severity (OR = 0.74, 95 percent CI: 0.40, 1.34), Qmax (OR = 1.11, 95 percent CI: 0.57, 2.15), prostate volume (OR = 1.17, 95 percent CI: 0.64, 2.17), and serum PSA level (OR = 0.74, 95 percent CI: 0.40, 1.38) were virtually the same as those for men whose serum IGFBP-3 levels were higher (OR = 1.01, 95 percent CI: 0.54, 1.88; OR = 1.13, 95 percent CI: 0.53, 2.42; OR = 1.40, 95 percent CI: 0.69, 2.90; and OR = 0.63, 95 percent CI: 0.33, 1.20, respectively). When we tested for an interaction between serum levels of IGFBP-3 and IGF-I, none of the associations differed by serum IGFBP-3 level.

DISCUSSION

In this study, we evaluated the cross-sectional association between serum levels of IGF-I and IGFBP-3 and surrogate markers of benign prostatic hyperplasia in a random sample of Caucasian, community-dwelling men. IGF-I has powerful mitogenic and antiapoptotic effects in many tissues (6, 7). These effects may be modulated by concentrations of IGFBP-3 in the peripheral circulation (21, 31), with increases in serum IGFBP-3 levels attenuating the mitogenic and antiapoptotic effects of IGF-I in tissues. Thus, we would expect to find a direct association between serum levels of IGF-I and our surrogate measures of benign prostatic hyperplasia and an inverse relation between serum levels of IGFBP-3 and these measures of benign prostatic hyperplasia.

For IGF-I, our cross-sectional data did not support these expected relations. Without adjustment for age, there was the suggestion of nonsignificant inverse associations between high IGF-I levels and urologic measures, a direction opposite to what was hypothesized. After adjustment for age, the point estimates for the associations between serum IGF-I and International Prostate Symptom Score, Qmax, and prostate volume approached 1, but the estimate for the association with serum PSA remained below 1. These nonsignificant findings could reflect the limits of cross-sectional data that provide a snapshot of inherently dynamic processes, could indicate that serum concentrations do not reflect intracellular levels where the biologic activity takes place (18, 21), or could indicate that there is no association.

By contrast, the cross-sectional associations between serum IGFBP-3 levels and urologic measures in this study were in the hypothesized direction, although they were of modest magnitude. Moreover, the unadjusted dose-response relation with prostate volume provided further credence to this relation. Although this finding is consistent with the hypothesis that more IGF-I may be bioavailable when IGFBP-3 levels are low, when IGFBP-3 was taken into account as either a confounder or an effect modifier, still no significant association was found between serum IGF-I levels and urologic measures. Adjustment for age did not dramatically alter these relations, but the statistical significance of the association between serum IGFBP-3 and serum PSA levels was lost. This finding suggests that the observed association may be a consequence of the confounding effects of age. In fact, when the analyses were stratified by age, we found no significant relation between serum IGFBP-3 levels and measures of benign prostatic hyperplasia within any of the age-specific strata (data not shown).

Other investigators have examined the associations between serum IGF-I or IGFBP-3 level and benign prostatic hyperplasia (11, 32, 33) or prostate size (12, 20). In general, the study findings were similar to ours in that the investigators observed no significant associations between serum IGF-I or serum IGFBP-3 level and benign prostatic hyperplasia (11, 12, 32). However, Chokkalingam et al. (33) observed a significant dose-response association between increasing IGFBP-3 level and prostate volume and a significant trend in risk of benign prostatic hyperplasia with increasing serum IGF-I level after adjustment for age. In our study, the dose-response association between IGFBP-3 and prostate volume was no longer evident after adjustment for age. Sarma et al. (20) reported an inverse correlation between IGF-I and prostate volume and a positive age-adjusted correlation between serum IGFBP-3 levels and prostate volume. The latter finding was in contrast to ours. The reported serum IGFBP-3 levels were similar to levels in Olmsted County men, but serum IGF-I levels were much lower; the reason is not apparent.

In a study of Finnish men that examined the association of prostate cancer with IGF-I and IGFBP-3, men with benign prostatic hyperplasia were used as controls (16). Reported serum IGF-I and IGFBP-3 levels and mean prostate volume for men with benign prostatic hyperplasia were higher than for men in our study and may suggest an association with benign prostatic hyperplasia (15), but no associations with urologic measures of benign prostatic hyperplasia were examined. The differences in findings across studies may relate to 1) variations in the definition and disease spectrum of benign prostatic hyperplasia: histologic benign prostatic hyperplasia (32, 33), prostate size (12, 20), or a more broad-spectrum definition based on surrogate measures, as in our study; 2) differences in assays for measuring serum IGF-I and IGFBP-3 levels (15, 34); or 3) racial differences in IGF-I and IGFBP-3 levels (35, 36). Although our study included US Caucasians, Stattin et al. (32) studied Swedes, Mantzoros et al. (11) studied Greeks, Sarma et al. (20) studied African-American men, and the Harman et al. (12) cohort was 87 percent Caucasian. The potential effects of racial differences on the metabolism of IGF-I within the prostate or the way in which IGF-I interacts with its binding proteins may be worth pursuing.

It is difficult to discern the meaning of our findings because of the complex interrelations between prostatic tissue and serum levels of IGF-I and IGFBP-3 and with intraprostatic androgens (5, 37). In a dynamic, in vivo system, it is difficult to evaluate the independent effects of each protein on prostatic growth. Through homeostatic mechanisms, changes in levels of one metabolite could result in compensatory responses. With cross-sectional data, the ability to assess temporal relations is lost.

Assessment of benign prostatic hyperplasia also complicates our ability to understand the associations with these proteins. Benign prostatic hyperplasia is a histologic diagnosis, but since it is not practical to perform a biopsy in order to make the diagnosis, it is necessary to rely on surrogate measures to define the condition. However, no single surrogate measure is a perfect marker for benign prostatic hyperplasia, and correlations among the surrogate measures are only modest (38). Each measure has a component related to benign prostatic hyperplasia and one due to other causes. Therefore, it is important to evaluate the association between exposure variables and more than one surrogate measure of benign prostatic hyperplasia. The suggestion of an association between serum IGFBP-3 levels and prostate volume, but not with serum PSA level (an indirect measure of epithelial volume) (39) or the other measures, may indicate that IGFBP-3 levels are associated with stromal proliferation.

Other potential limitations should be noted when interpreting our results. With each of the urologic measures, measurement error (random variability) could have reduced our ability to detect a correlation or an association. Although the participation rate of 55 percent for the entire cohort suggests a potential for bias, it seems unlikely that participants would differ from nonparticipants in outcomes conditional on exposure (IGF-I or IGFBP-3). A review of the medical records of participants and nonparticipants, after approval by the Mayo Institutional Review Board, indicated a slightly older age and a higher prevalence of urologic diagnoses among participants (40) that did not appear to bias long-term outcomes (30). Finally, our findings are generalizable primarily to Caucasian men, and extrapolation of these findings to men not represented in the study should be performed with caution.

In summary, our data failed to demonstrate significant associations between serum levels of IGF-I and surrogate measures of benign prostatic hyperplasia. However, they suggest that low serum levels of IGFBP-3 may be associated with increased prostate size, although this possibility may just reflect confounding by age.

ACKNOWLEDGMENTS

This project was supported by research grants from the Public Health Service, National Institutes of Health (DK58859 and AR30582), and Merck Research Laboratories.

The authors thank the entire benign prostatic hyperplasia staff for their help with the study and Sondra Buehler for her adept preparation of the manuscript.

Reprint requests to Dr. Rosebud Roberts, Department of Health Sciences Research, Mayo Clinic, 200 First Street S.W., Rochester, MN 55905 (e-mail: roberts.rosebud@mayo.edu).

TABLE 1.

Distribution of age and urologic measures of benign prostatic hyperplasia and their correlations with serum IGF-I† and IGFBP-3† levels, Olmsted County Study, Minnesota, 1990–1996

 No. Mean (SD†) Correlation coefficients 
 IGF-I  IGFBP-3 
 Spearman Partial‡  Spearman Partial‡ 
Age (years) 451 59.2 (10.5) –0.21***   –0.33***  
IPSS† 438 8.0 (6.0) –0.07 –0.03  –0.12** –0.04 
Qmax† (ml/second) 434 19.4 (10.2) 0.004 –0.06  0.10* 0.01 
Prostate volume (ml) 402 29.9 (12.7) –0.08 –0.03  –0.14** 0.003 
Serum PSA† (ng/ml) 451 1.4 (0.97) –0.15** –0.06  –0.14** 0.02 
Serum IGF-I (ng/ml) 450 134.1 (55.3) 1.0 1.0  0.61*** 0.59*** 
Serum IGFBP-3 (ng/ml) 450 3,271.6 (650.9) 0.61*** 0.59***  1.0 1.00 
 No. Mean (SD†) Correlation coefficients 
 IGF-I  IGFBP-3 
 Spearman Partial‡  Spearman Partial‡ 
Age (years) 451 59.2 (10.5) –0.21***   –0.33***  
IPSS† 438 8.0 (6.0) –0.07 –0.03  –0.12** –0.04 
Qmax† (ml/second) 434 19.4 (10.2) 0.004 –0.06  0.10* 0.01 
Prostate volume (ml) 402 29.9 (12.7) –0.08 –0.03  –0.14** 0.003 
Serum PSA† (ng/ml) 451 1.4 (0.97) –0.15** –0.06  –0.14** 0.02 
Serum IGF-I (ng/ml) 450 134.1 (55.3) 1.0 1.0  0.61*** 0.59*** 
Serum IGFBP-3 (ng/ml) 450 3,271.6 (650.9) 0.61*** 0.59***  1.0 1.00 

* p < 0.05; ** p < 0.01; *** p < 0.001.

† IGF-I, insulin-like growth factor I; IGFBP-3, insulin-like growth factor binding protein 3; SD, standard deviation; IPSS, International Prostate Symptom Score; Qmax, peak urinary flow rate; PSA, prostate-specific antigen.

‡ Partial correlations were adjusted for age.

TABLE 2.

Distribution of serum IGF-I* and IGFBP-3* levels by age and urologic measures, Olmsted County Study, Minnesota, 1990–1996

 No. Serum IGF-I (ng/ml)  Serum IGFBP-3 (ng/ml) 
Mean (SD*) p value†  Mean (SD) p value† 
Age (years)   0.0002   0.0001 
40–49 114 150.1 (62.4)   3,461.5 (626.3)  
50–59 107 137.7 (51.0)   3,453.2 (657.8)  
60–69 83 123.7 (52.1)   3,090.1 (520.1)  
≥70 146 119.3 (50.3)   2,925.5 (625.7)  
IPSS*   0.07   0.05 
≤7 228 138.6 (57.9)   3,329.6 (646.5)  
>7 210 129.1 (51.2)   3,207.4 (654.6)  
Qmax* (ml/second)   0.5   0.02 
≥12 321 135.3 (57.6)   3,321.0 (658.5)  
<12 113 131.0 (49.9)   3,153.5 (617.1)  
Prostate volume (ml)   0.26   0.003 
≤30 239 136.8 (53.8)   3,356.3 (638.5)  
>30 163 130.4 (58.9)   3,156.8 (667.9)  
Serum PSA* (ng/ml)   0.002   0.003 
≤1.4 307 140.2 (58.8)   3,330.1 (664.0)  
>1.4 144 120.7 (46.3)   3,152.7 (626.5)  
 No. Serum IGF-I (ng/ml)  Serum IGFBP-3 (ng/ml) 
Mean (SD*) p value†  Mean (SD) p value† 
Age (years)   0.0002   0.0001 
40–49 114 150.1 (62.4)   3,461.5 (626.3)  
50–59 107 137.7 (51.0)   3,453.2 (657.8)  
60–69 83 123.7 (52.1)   3,090.1 (520.1)  
≥70 146 119.3 (50.3)   2,925.5 (625.7)  
IPSS*   0.07   0.05 
≤7 228 138.6 (57.9)   3,329.6 (646.5)  
>7 210 129.1 (51.2)   3,207.4 (654.6)  
Qmax* (ml/second)   0.5   0.02 
≥12 321 135.3 (57.6)   3,321.0 (658.5)  
<12 113 131.0 (49.9)   3,153.5 (617.1)  
Prostate volume (ml)   0.26   0.003 
≤30 239 136.8 (53.8)   3,356.3 (638.5)  
>30 163 130.4 (58.9)   3,156.8 (667.9)  
Serum PSA* (ng/ml)   0.002   0.003 
≤1.4 307 140.2 (58.8)   3,330.1 (664.0)  
>1.4 144 120.7 (46.3)   3,152.7 (626.5)  

* IGF-I, insulin-like growth factor I; IGFBP-3, insulin-like growth factor binding protein 3; SD, standard deviation; IPSS, International Prostate Symptom Score; Qmax, peak urinary flow rate; PSA, prostate-specific antigen.

† Nonparametric Wilcoxon rank sum test and Kruskall-Wallis test (age).

TABLE 3.

Crude and age-adjusted associations between serum IGF-I* levels and urologic measures, Olmsted County Study, Minnesota, 1990–1996

Variable IPSS* >7  Qmax* <12 ml/second  Prostate volume >30 ml  PSA* >1.4 ng/ml 
OR* 95% CI*  OR 95% CI  OR 95% CI  OR 95% CI 
Unadjusted for age            
IGF-I >median 0.81 0.55, 1.17  0.89 0.58, 1.37  0.88 0.59, 1.32  0.58 0.39, 0.86 
IGF-I ≤median† 1.0   1.0   1.0   1.0  
Age adjusted‡            
IGF-I >median 0.98 0.66, 1.45  1.14 0.72, 1.80  1.11 0.72, 1.72  0.71 0.46, 1.09 
IGF-I ≤median 1.0   1.0   1.0   1.0  
Variable IPSS* >7  Qmax* <12 ml/second  Prostate volume >30 ml  PSA* >1.4 ng/ml 
OR* 95% CI*  OR 95% CI  OR 95% CI  OR 95% CI 
Unadjusted for age            
IGF-I >median 0.81 0.55, 1.17  0.89 0.58, 1.37  0.88 0.59, 1.32  0.58 0.39, 0.86 
IGF-I ≤median† 1.0   1.0   1.0   1.0  
Age adjusted‡            
IGF-I >median 0.98 0.66, 1.45  1.14 0.72, 1.80  1.11 0.72, 1.72  0.71 0.46, 1.09 
IGF-I ≤median 1.0   1.0   1.0   1.0  

* IGF-I, insulin-like growth factor I; IPSS, International Prostate Symptom Score; Qmax, peak urinary flow rate; PSA, prostate-specific antigen; OR, odds ratio; CI, confidence interval.

† Reference group: IGF-I level ≤132 ng/ml (median); IGF-I levels were >median for 299 men and ≤median for 221 men.

‡ Adjusted for 10-year age groups.

TABLE 4.

Crude and age-adjusted associations between serum IGFBP-3* levels and urologic measures, Olmsted County Study, Minnesota, 1990–1996

Variable IPSS* >7  Qmax* <12 ml/second  Prostate volume >30 ml  PSA* >1.4 ng/ml 
OR* 95% CI*  OR 95% CI  OR 95% CI  OR 95% CI 
Unadjusted for age            
IGFBP-3 <median 1.22 0.84, 1.78  1.48 0.96, 2.28  1.90 1.27, 2.89  1.71 1.14, 2.55 
IGFBP-3 ≥median† 1.0   1.0   1.0   1.0  
Age adjusted‡            
IGFBP-3 <median 0.89 0.59, 1.34  1.06 0.67, 1.69  1.46 0.94, 2.27  1.19 0.77, 1.86 
IGFBP-3 ≥median 1.0   1.0   1.0   1.0  
Variable IPSS* >7  Qmax* <12 ml/second  Prostate volume >30 ml  PSA* >1.4 ng/ml 
OR* 95% CI*  OR 95% CI  OR 95% CI  OR 95% CI 
Unadjusted for age            
IGFBP-3 <median 1.22 0.84, 1.78  1.48 0.96, 2.28  1.90 1.27, 2.89  1.71 1.14, 2.55 
IGFBP-3 ≥median† 1.0   1.0   1.0   1.0  
Age adjusted‡            
IGFBP-3 <median 0.89 0.59, 1.34  1.06 0.67, 1.69  1.46 0.94, 2.27  1.19 0.77, 1.86 
IGFBP-3 ≥median 1.0   1.0   1.0   1.0  

* IGFBP-3, insulin-like growth factor binding protein 3; IPSS, International Prostate Symptom Score; Qmax, peak urinary flow rate; PSA, prostate-specific antigen; OR, odds ratio; CI, confidence interval.

† Reference group: IGFBP-3 level ≥3,227 ng/ml (median). For equal numbers of men (n = 225), IGFBP-3 levels were ≥median and <median.

‡ Adjusted for 10-year age groups.

TABLE 5.

Multivariable associations between urologic measures and serum IGFBP-3* and IGF-I* levels, with simultaneous adjustment for IGF-I and IGFBP-3 with and without adjustment for age, Olmsted County Study, Minnesota, 1990–1996

Variable IPSS* >7  Qmax* <12 ml/second  Prostate volume >30 ml  PSA* >1.4 ng/ml 
OR* 95% CI*  OR 95% CI  OR 95% CI  OR 95% CI 
Unadjusted for age            
IGFBP-3 <median 1.13 0.74, 1.74  1.57 0.95, 2.59  2.14 1.34, 3.40  1.43 0.91, 2.25 
IGFBP-3 ≥median† 1.0   1.0   1.0   1.0  
IGF-1 >median 0.86 0.56, 1.32  1.12 0.68, 1.84  1.27 0.80, 2.02  0.69 0.44, 1.08 
IGF-1 ≤median‡ 1.0   1.0   1.0   1.0  
Age adjusted§            
IGFBP-3 <median 0.85 0.53, 1.35  1.19 0.69, 2.04  1.72 1.05, 2.82  1.02 0.62, 1.68 
IGFBP-3 ≥median 1.0   1.0   1.0   1.0  
IGF-1 >median 0.90 0.58, 1.42  1.24 0.73, 2.11  1.42 0.87, 2.33  0.72 0.44, 1.16 
IGF-1 ≤median 1.0   1.0   1.0   1.0  
Variable IPSS* >7  Qmax* <12 ml/second  Prostate volume >30 ml  PSA* >1.4 ng/ml 
OR* 95% CI*  OR 95% CI  OR 95% CI  OR 95% CI 
Unadjusted for age            
IGFBP-3 <median 1.13 0.74, 1.74  1.57 0.95, 2.59  2.14 1.34, 3.40  1.43 0.91, 2.25 
IGFBP-3 ≥median† 1.0   1.0   1.0   1.0  
IGF-1 >median 0.86 0.56, 1.32  1.12 0.68, 1.84  1.27 0.80, 2.02  0.69 0.44, 1.08 
IGF-1 ≤median‡ 1.0   1.0   1.0   1.0  
Age adjusted§            
IGFBP-3 <median 0.85 0.53, 1.35  1.19 0.69, 2.04  1.72 1.05, 2.82  1.02 0.62, 1.68 
IGFBP-3 ≥median 1.0   1.0   1.0   1.0  
IGF-1 >median 0.90 0.58, 1.42  1.24 0.73, 2.11  1.42 0.87, 2.33  0.72 0.44, 1.16 
IGF-1 ≤median 1.0   1.0   1.0   1.0  

* IGFBP-3, insulin-like growth factor binding protein 3; IGF-I, insulin-like growth factor I; IPSS, International Prostate Symptom Score; Qmax, peak urinary flow rate; PSA, prostate-specific antigen; OR, odds ratio; CI, confidence interval.

† Reference group: IGFBP-3 level ≥3,227 ng/ml (median).

‡ Reference group: IGF-I level ≤132 ng/ml (median).

§ Adjusted for 10-year age groups.

References

1.
Serio M, Fiorelli G. Dual control by androgens and peptide growth factors of prostatic growth in human benign prostatic hyperplasia.
Mol Cell Endocrinol
 
1991
;
78
:
C77
–81.
2.
Culig Z, Hobisch A, Cronauer MV, et al. Regulation of prostatic growth and function by peptide growth factors.
Prostate
 
1996
;
28
:
392
–405.
3.
Cohen P, Peehl DM, Baker B, et al. Insulin-like growth factor axis abnormalities in prostatic stromal cells from patients with benign prostatic hyperplasia.
J Clin Endocrinol Metab
 
1994
;
79
:
1410
–15.
4.
Boudon C, Rodier G, Lechevallier E, et al. Secretion of insulin-like growth factors and their binding proteins by human normal and hyperplastic prostatic cells in primary culture.
J Clin Endocrinol Metab
 
1996
;
81
:
612
–17.
5.
Monti S, Di Silverio F, Iraci R, et al. Regional variations of insulin-like growth factor I (IGF-I), IGF-II, and receptor type I in benign prostatic hyperplasia tissue and their correlation with intraprostatic androgens.
J Clin Endocrinol Metab
 
2001
;
86
:
1700
–6.
6.
Grant ES, Ross MB, Ballard S, et al. The insulin-like growth factor type I receptor stimulates growth and suppresses apoptosis in prostatic stromal cells.
J Clin Endocrinol Metab
 
1998
;
83
:
3252
–7.
7.
Cory S, Vaux DL, Strasser A, et al. Insights from Bcl-2 and Myc: malignancy involves abrogation of apoptosis as well as sustained proliferation.
Cancer Res
 
1999
;
59
:
1685s
–92s.
8.
Chan JM, Stampfer MJ, Giovannucci E, et al. Plasma insulin-like growth factor-I and prostate cancer risk: a prospective study.
Science
 
1998
;
279
:
563
–6.
9.
Wolk A, Mantzoros CS, Andersson SO, et al. Insulin-like growth factor 1 and prostate cancer risk: a population-based, case-control study.
J Natl Cancer Inst
 
1998
;
90
:
911
–15.
10.
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
–17.
11.
Mantzoros CS, Tzonou A, Signorello LB, et al. Insulin-like growth factor 1 in relation to prostate cancer and benign prostatic hyperplasia.
Br J Cancer
 
1997
;
76
:
1115
–18.
12.
Harman SM, Metter EJ, Blackman MR, et al. Serum levels of insulin-like growth factor I (IGF-I), IGF-II, IGF-binding protein-3, and prostate-specific antigen as predictors of clinical prostate cancer.
J Clin Endocrinol Metab
 
2000
;
85
:
4258
–65.
13.
Colao A, Marzullo P, Ferone D, et al. Prostatic hyperplasia: an unknown feature of acromegaly.
J Clin Endocrinol Metab
 
1998
;
83
:
775
–9.
14.
Colao A, Marzullo P, Spiezia S, et al. Effect of growth hormone (GH) and insulin-like growth factor I on prostate diseases: an ultrasonographic and endocrine study in acromegaly, GH deficiency, and healthy subjects.
J Clin Endocrinol Metab
 
1999
;
84
:
1986
–91.
15.
Cohen P. Serum insulin-like growth factor-I levels and prostate cancer risk—interpreting the evidence.
J Natl Cancer Inst
 
1998
;
90
:
876
–9.
16.
Finne P, Auvinen A, Koistinen H, et al. Insulin-like growth factor I is not a useful marker of prostate cancer in men with elevated levels of prostate-specific antigen.
J Clin Endocrinol Metab
 
2000
;
85
:
2744
–7.
17.
Rajaram S, Baylink DJ, Mohan S. Insulin-like growth factor-binding proteins in serum and other biological fluids: regulation and functions.
Endocr Rev
 
1997
;
18
:
801
–31.
18.
Djavan B, Waldert M, Seitz C, et al. Insulin-like growth factors and prostate cancer.
World J Urol
 
2001
;
19
:
225
–33.
19.
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.
20.
Sarma AV, Jaffe CA, Schottenfeld D, et al. Insulin-like growth factor-1, insulin-like growth factor binding protein-3, and body mass index: clinical correlates of prostate volume among black men.
Urology
 
2002
;
59
:
362
–7.
21.
Huynh H, Seyam RM, Brock GB. Reduction of ventral prostate weight by finasteride is associated with suppression of insulin-like growth factor I (IGF-I) and IGF-I receptor genes and with an increase in IGF binding protein 3.
Cancer Res
 
1998
;
58
:
215
–18.
22.
Jacobsen SJ, Guess HA, Panser L, et al. A population-based study of health care-seeking behavior for treatment of urinary symptoms: The Olmsted County Study of Urinary Symptoms and Health Status Among Men.
Arch Fam Med
 
1993
;
2
:
729
–35.
23.
Chute CG, Panser LA, Girman CJ, et al. The prevalence of prostatism: a population-based survey of urinary symptoms.
J Urol
 
1993
;
150
:
85
–9.
24.
Kurland LT, Molgaard CA. The patient record in epidemiology.
Sci Am
 
1981
;
245
:
54
–63.
25.
Melton LJ III. History of the Rochester Epidemiology Project.
Mayo Clin Proc
 
1996
;
71
:
266
–74.
26.
Oesterling JE, Jacobsen SJ, Chute CG, et al. Serum prostate-specific antigen in a community-based population of healthy men: establishment of age-specific reference ranges.
JAMA
 
1993
;
270
:
860
–4.
27.
Jacobsen SJ, Girman CJ, Guess HA, et al. Natural history of prostatism: longitudinal changes in voiding symptoms in community-dwelling men.
J Urol
 
1996
;
155
:
595
–600.
28.
Barry MJ, Fowler FJJ, O’Leary MP, et al. The American Urological Association Symptom Index for Benign Prostatic Hyperplasia.
J Urol
 
1992
;
148
:
1549
–57.
29.
Terris MK, Stamey TA. Determination of prostate volume by transrectal ultrasound.
J Urol
 
1991
;
145
:
984
–7.
30.
Jacobsen SJ, Jacobson DJ, Girman CJ, et al. Natural history of prostatism: risk factors for acute urinary retention.
J Urol
 
1997
;
158
:
481
–7.
31.
Sutkowski DM, Goode RL, Baniel J, et al. Growth regulation of prostatic stromal cells by prostate-specific antigen.
J Natl Cancer Inst
 
1999
;
91
:
1663
–9.
32.
Stattin P, Kaaks R, Riboli E, et al. Circulating insulin-like growth factor-I and benign prostatic hyperplasia: a prospective study.
Scand J Urol Nephrol
 
2001
;
35
:
122
–6.
33.
Chokkalingam AP, Gao YT, Deng J, et al. Insulin-like growth factors and risk of benign prostatic hyperplasia.
Prostate
 
2002
;
51
:
98
–105.
34.
Cohen P, Peehl DM, Rosenfeld R, et al. Insulin-like growth factor 1 in relation to prostate cancer and benign prostatic hyperplasia.
Br J Cancer
 
1998
;
78
:
554
–6.
35.
Platz EA, Pollak MN, Rimm EB, et al. Racial variation in insulin-like growth factor-1 and binding protein-3 concentrations in middle-aged men.
Cancer Epidemiol Biomarkers Prev
 
1999
;
8
:
1107
–10.
36.
Tricoli JV, Winter DL, Hanlon AL, et al. Racial differences in insulin-like growth factor binding protein-3 in men at increased risk of prostate cancer.
Urology
 
1999
;
54
:
178
–82.
37.
Monti S, Di Silverio F, Lanzara S, et al. Insulin-like growth factor-I and -II in human benign prostatic hyperplasia: relationship with binding proteins 2 and 3 and androgens.
Steroids
 
1998
;
63
:
362
–6.
38.
Girman CJ, Jacobsen SJ, Guess HA, et al. Natural history of prostatism: relationship among symptoms, prostate volume and peak urinary flow rate.
J Urol
 
1995
;
153
:
1510
–15.
39.
Weber JP, Oesterling JE, Peters CA, et al. The influence of reversible androgen deprivation on serum prostate-specific antigen levels in men with benign prostatic hyperplasia.
J Urol
 
1989
;
141
:
987
–92.
40.
Panser LA, Chute CG, Guess HA, et al. The natural history of prostatism: the effects of non-response bias.
Int J Epidemiol
 
1994
;
23
:
1198
–205.