SERUM TOTAL prostate-specific antigen (PSA) concentrations

Transcription

1 X/00/$03.00/0 Vol. 85, No. 3 The Journal of Clinical Endocrinology & Metabolism Printed in U.S.A. Copyright 2000 by The Endocrine Society Heritability of Prostate-Specific Antigen and Relationship with Zonal Prostate Volumes in Aging Twins* ARUNA BANSAL, DARRELL K. MURRAY, JAMES T. WU, ROBERT A. STEPHENSON, RICHARD G. MIDDLETON, AND A. WAYNE MEIKLE Departments of Medicine (D.K.M., A.W.M.), Urology (R.A.S., R.G.M.), Pathology (J.T.W., A.W.M.), and Medical Informatics (A.B.), University of Utah School of Medicine and the ARUP Institute, Salt Lake City, Utah ABSTRACT Both benign prostatic hyperplasia and prostate-specific antigen (PSA) have been shown to increase with age and with prostate volume in men, but the influence of heredity on these relationships is not completely understood. This study has two aims: 1) to investigate the inter-relationships of age, PSA, and various zonal measurements in the prostate; and 2) to assess the impact of heritable influences on total PSA. Eighty-four monozygotic twin pairs and 83 dizygotic twin pairs were studied, and serum total PSA, free PSA, and PSA- 1 - antichymotrypsin were measured. Their prostate volumes [total (TV), transition zone (TZ), and peripheral zone) were quantitated using transrectal ultrasound. SERUM TOTAL prostate-specific antigen (PSA) concentrations are influenced by many factors, including aging, prostate volume, prostate cancer, and prostatitis (1). PSA is a glycoprotein with serine protease activity that is synthesized mainly in the ductal epithelium and prostatic acini and is encoded by a gene on the long arm of chromosome 19 (2 4). Seminal plasma has a concentration about a million times higher than serum, where most of the PSA is bound to a protease inhibitor 1 -antichymotrypsin (ACT). A smaller proportion is bound to 2 -macroglobulin, and a small fraction is free (5, 6). Although much is known about factors that affect serum total PSA, the degree of hereditary influence on variation in serum total PSA has not been reported, to our knowledge. Several methods have been recommended to improve the use of serum total PSA in the diagnosis of prostate cancer, and one is the use of an age-specific reference range (1, 7 9). Whereas age-related prostate enlargement frequently results in benign prostatic hyperplasia (BPH) as a consequence of a greater growth of the transition zone (TZ) than the peripheral zone (PZ) (10 12), both zones can affect serum total PSA values. Roehrborn et al. (13, 14) concluded that prostate volume (TV) is strongly related to serum total PSA in men with Received August 12, Revision received November 2, Accepted November 9, Address correspondence and requests for reprints to: A. Wayne Meikle, M.D., University of Utah School of Medicine, 50 North Medical Drive, Salt Lake City, Utah * Supported in part by NIH USPHS Grants DK-45760, DK-43344, and RR Total PSA is significantly correlated with all zonal prostate measurements (TZ, peripheral zone, TV, and TZ/TV) and with age. When linear regression was applied, only age and TZ were retained in the final model. The proportion of variability in total PSA explained by these two factors, however, is below 24%. In contrast, estimates of heritability show that approximately 45% of the variability in total PSA can be explained by inherited factors. Whereas age and TZ are linearly related to total PSA, their influence is much less than that of familial and genetic factors. It is uncertain whether these factors predispose also to prostate cancer or if they are independent of those, whether they confound the accuracy of using total serum PSA level as a diagnostic tool. (J Clin Endocrinol Metab 85: , 2000) BPH but without prostate cancer and that age affects the relationship between TV and total PSA (13, 15). PSA density [total serum PSA (ng/ml)/prostate volume (cm 3 )] has been proposed to differentiate men with BPH from those who harbor prostate cancer. It has been shown that when the density exceeds 0.15, prostate cancer is more likely (13, 14, 16 24). PSA velocity is another useful indicator for occult prostate cancer (25, 26). When the velocity was 0.75 ng/ml/yr or greater, prostate cancer was more likely, and prostate biopsy has been recommended (25, 26). The ratio of free PSA to bound PSA has also been recommended for selecting men with possible prostate cancer. If the ratio is more than 25%, the likelihood of prostate cancer was less than 10% compared to a risk more than 80% if the ratio were less than 10% (17, 27 32). In the current study, we report on the inter-relationships of age, zonal prostate volumes, and heredity on PSA levels in male twins ranging in age from yr. Materials and Methods A population-based male twin registry was developed from database registries, and it has had a recent update of driver s license information (10, 11). Genealogy information has been linked to birth and death records and driver s license data. About 95% of individuals in the registry are Caucasian. The twin pairs were selected at random and were invited to participate if they were between 25 and 77 yr of age and both twin pairs were living. They were not preselected based on PSA values. We report data on 84 pairs of monozygotic (MZ) twins (age, yr; range, yr) and 83 pairs of dizygotic (DZ) twins (age, yr; range, yr), for which both members of the pair were available for evaluation and had no endocrine or other health issues. All men with prostate cancer or a history of treatment for BPH were excluded from analysis. Informed consent was obtained before entry into the study, 1272

2 AGE, PSA, AND PROSTATE ZONES 1273 which was conducted in the Clinical Research Center of the University of Utah. None of the subjects was receiving medications known to affect PSA values. Following a digital rectal examination, transrectal ultrasound was performed using a Bruel and Kjaer 1 instrument (Cincinnati, OH) fitting with a 7 MHz. transducer. The total TV (assuming 1 cc 1 g) and the volume of the TZ were calculated by /6 times width (maximal transverse dimension), length (maximal anterior and posterior dimension), and height (maximal sagittal proximal to distal dimension) (33). Because the TZ is significantly hypoechoic relative to PZ echodensities, accurate visualization of zonal boundaries is possible for the purposes of measurements used in volume calculations. The PZ volume (central zone is included in PZ calculation) was estimated by subtracting the TZ from the TV. The coefficient of variation was 5% for TV and 11% for TZ. PSA determinations Between 0800 and 1030 h, three blood samples were obtained in tubes by venipuncture at intervals of min for the determination of PSA concentration. The same volume from each specimen was pooled and stored at 20 C until assayed. PSA was determined with the Hybritech Tandem E assay (12, 34, 35). PSA-ACT, ACT, and free PSA were measured by the method of Wu et al. (36). Statistical analyses Statistical analysis was undertaken to determine the degree of linear relationship among the 10 variables of interest. These consisted of total PSA, free PSA, PSA-ACT, and sum of free PSA and PSA-ACT, TV, PZ, TZ, the ratio of TZ/TV, and age. Table 1 shows the Pearson Product Moment correlation coefficient (r) for all pairs of variables. At this stage in the analysis, only the first twin of each pair was used, to avoid the bias of using individuals who are clearly correlated. As the table shows, sample size for some PSA measurements falls below 60. These have, therefore, not been used in linear regression analysis (below). Concentrating on total PSA as the outcome of interest, r 2 -values were calculated to determine what proportion of the total sum of squares of total PSA could be explained by the variables age, TZ volume, PZ volume, TV volume, and the TZ/TV ratio. A histogram of total PSA showed it to be positively skewed, so a logarithmic transformation was applied to enhance linearity and homogeneity of variance in preparation for linear regression. Using age, TZ, PZ, TV, and TZ/TV ratio as independent variables in the full model, stepwise linear regression was applied to the log-transformed total PSA. The most parsimonious model is shown in Table 3. The issues of heritability of total PSA and of PSA density [total serum PSA (ng/ml)/prostate volume (cm 3 )] were also evaluated. Both members of each twin pair were incorporated into this stage of the analysis. Heritability (h 2 ) represents that part of the phenotypic variance that is accounted for by additive genetic factors and is obtained by the formula h 2 2(r MZ r DZ ) where r MZ and r DZ denote the intraclass correlation values for MZ and DZ twins, respectively (37). Because the importance of heritable factors can vary at different ages, h 2 was calculated in two ways: first, using all twins, and then using twins stratified by age. For this second step, a cut-off of age 50 yr was used, leading to sample sizes of 17 MZ and 23 DZ twin pairs below age 50 yr and 45 MZ and 25 DZ twin pairs aged 50 yr or above. It was acknowledged that the small TABLE 1. Pearson s product-moment correlation coefficients (r) sample sizes make the age-stratified estimates less reliable than the nonstratified estimate. Results In a two-tailed test of r 0, age was significantly ( 0.05) correlated with total PSA, free PSA, and all of the zonal measurements. As expected, in every case, the correlation was positive. The correlation of age with PSA-ACT, sum of PSA-ACT and free PSA and free/psa was, in each case, positive but nonsignificant. Not surprisingly, significant positive correlation was found among the four linear PSA measurements: total, free, PSA-ACT, and summed PSA. The same was seen among PZ, TZ, and TV. Total PSA showed a significant positive linear correlation with all three zonal measurements, and, in addition, with the TZ/TV ratio. All of these variables were, therefore, used in the full model for stepwise linear regression (below). Free PSA behaved in a very similar way, although the correlation coefficient for TZ/TV was nonsignificant. Interestingly, PSA-ACT gave nonsignificant correlation with all zonal measurements, whereas free/psa was only significantly correlated with PZ. Summed PSA was significantly correlated with both TZ and TV. Focusing on total PSA, stepwise linear regression was performed using the independent variables: age, PZ, TZ, TV and TZ/TV (summary statistics are given in Table 2). These had r 2 values of 0.16, 0.03, 0.13, 0.08, and 0.05, respectively, showing the clear importance of age and TZ. This was reflected also in the regression results. Graphs of these two variables plotted against ln (total PSA) are given in Fig. 1. This was also reflected in the regression results. The most parsimonious model is described in Table 3, and it consisted of age, TZ, and an intercept term. All other zonal measurements were nonsignificant at the 5% level once age and TZ had been included. To test the hypothesis that inherited factors play a role in PSA levels, heritability (h 2 ) was estimated using intraclass correlation coefficients derived from the twin data (Table 4). It emerged that the heritability estimates for total PSA were hardly affected by age-stratification. Using data at all ages, h 2 was For young twins ( 50 yr), it was , and for older twins ( 50 yr) it was In contrast, in looking at PSA density, the h 2 for all twins was 0.33; for those younger than age 50, it was 0.195, and for those age 50 and over, it was n Total PSA Free PSA PSA-ACT Free/PSA Sum PSA PZ TZ TV TZ/TV Age Total PSA Free PSA PSA-ACT Free/PSA Sum PSA PZ TZ TV Sum, PSA-ACT plus free PSA.

3 1274 BANSAL ET AL. JCE&M 2000 Vol 85 No 3 TABLE 2. Summary of variables used in the linear regression Total PSA Age (yr) PZ TZ TV Minimum Mean Median Max SD FIG. 1. ln (Total PSA)-plotted age and TZ volume. TABLE 3. Derived model of log (total PSA) Value SE T-value P Intercept Age TZ TABLE 4. Heritability estimates of PSA Measure Age 50 yr Age 50 yr All ages Total PSA PSA density Discussion Our results are consistent with those reported recently by Roehrborn et al. (13), who observed that TV is strongly related to serum total PSA in aging men with BPH without prostate cancer. Age has also previously been shown to be important. The steeper rise in total PSA in older men without prostate cancer has lead to the recommendation of an ageadjusted reference range for PSA (7, 8, 13, 16, 30). Our observations confirm previous findings and extend them to show that heredity affects total serum PSA more than prostate zonal volumes or age. Whereas PSA-ACT and free PSA may offer advantages in limiting the need for biopsies in men with elevated PSA values suspected of harboring occult prostate cancer, these measurements were closely correlated with total PSA and were similar in terms of their relationships with prostate zonal volumes and age. All values of the correlation coefficient were positive and not significantly different. Total PSA showed a higher correlation with age and TZ than free PSA or PSA-ACT, but PSA-ACT was most strongly correlated with TZ of the three volume measurements, suggesting that sample size differences may have been responsible for the quantitative discrepancy. In contrast, free PSA showed higher correlation with PZ and TV than total PSA. A higher ratio of free PSA to PSA might be expected if the TZ contributed more than the PZ, where prostate cancer is likely to originate. In men without prostate cancer, however, this ratio shows no preferential relationship for the TZ compared with the PZ (17, 27 32). Linear regression suggested that age and TZ were most influential in determining total PSA levels. Because the three volume measurements are strongly correlated, it is not surprising that only one appeared in the final model. What is suggested, however, is that total PSA is more

4 AGE, PSA, AND PROSTATE ZONES 1275 strongly correlated with TZ volume than either PZ or TV. This is consistent with previous results showing that although the enlargement of the PZ and TZ occurring after puberty can contribute to a rise in total PSA, growth of the TZ is the main contributor to TV enlargement in men after age 50 (10, 11). Other studies have explored these relationships. The ageassociation with total PSA has been observed in men without detectable prostate cancer, and it seems to be dependent on prostate enlargement from benign disease (7 9, 13, 15, 18). Prostate volume and serum total PSA have been reported to exhibit an age-dependent log-linear relationship, suggesting that their logarithms are linearly related and age dependent (13), as we also have seen. The rise in TV tends to be higher in older men, leading to increases in serum total PSA (PSA, 1.6 ng/ml, 2.0 ng/ml, and 2.3 ng/ml for men with BPH in their 50s, 60s, and 70s, respectively) (13). Our observations of the correlation of age and TZ with serum total PSA are consistent with these previous reports. Although our model provided the best fit from the available variables, its r 2 value was relatively low, only This suggested that much of the variation in total PSA remained unexplained after taking into account age and TV. For the first time, we present the importance of heritability in total serum PSA levels. At all ages studied, at least 40% of the variation in total PSA can be attributed to inherited (both familial and genetic) effects. The current study suggests that these play a more significant role than either age or zonal prostate measurements. Our results demonstrate the importance of inherited factors on total PSA levels and show no strong evidence of differences in heritability at different ages. In contrast, there is some evidence that the heritability of PSA density may be higher in older than in younger men. Because TZ is correlated with total PSA and more profoundly affects TV than PZ in aging men, this may be indirectly related to hereditary influences working independently on total PSA and TZ. The latter has previously been shown to have approximately 25% of the variation influenced by hereditary factors (11). Serum total PSA has become an important marker for the presence of overt and occult prostate cancer. Because the variation of serum total PSA concentrations is strongly influenced by genetic factors, it is important to determine the mode of inheritance of serum total PSA. The disease has been reported to have a heritability of 57% (38). It is unclear whether men with an inherited tendency to higher serum total PSA concentrations are at higher risk to develop prostate cancer. However, individuals with a high PSA certainly undergo more intensive screening, including prostate biopsies. Such intensive screening procedures might lead to an apparently higher diagnostic rate of prostate cancer in them. Conversely, men with an inherited tendency to lower total PSA concentrations may go without biopsy inappropriately for prostate cancer. At least 22% of clinical prostate cancer occurs in men with a serum PSA measurement less than 4 ng/ml (31). An understanding of the genetic basis of serum total PSA levels is essential to allow greater precision in the interpretation of total PSA levels in a given patient. Further study is needed to determine the impact of heritability of PSA on detection of prostate cancer. References 1. Chan D, Sokoll L Prostate-specific antigen: update J Int Fed Clin Chem. 9: Lundwall A, Lilja H Molecular cloning of human prostate specific antigen cdna. FEBS Lett. 214: Watt KW, Lee PJ, M Timkulu T, Chan WP, Loor R Human prostatespecific antigen: structural and functional similarity with serine proteases. Proc Natl Acad Sci USA. 83: Riegman PH, Vlietstra RJ, Suurmeijer L, Cleutjens CB, Trapman J Characterization of the human kallikrein locus. Genomics. 14: Christensson A, Laurell CB, Lilja H Enzymatic activity of prostatespecific antigen and its reactions with extracellular serine proteinase inhibitors. Eur J Biochem. 194: Vessella RL, Lange PH Issues in the assessment of PSA immunoassays. Urol Clin North Am. 20: Morgan TO, Jacobsen SJ, McCarthy WF, Jacobson DJ, McLeod DG, Moul JW Age-specific reference ranges for prostate-specific antigen in black men. N Engl J Med. 335: Oesterling JE, Cooner WH, Jacobsen SJ, Guess HA, Lieber MM Influence of patient age on the serum PSA concentration. An important clinical observation. Urol Clin North Am. 20: 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 [see comments]. J Am Med Assoc. 270: Meikle AW, Stephenson RA, Lewis CM, Middleton RG Effects of age and sex hormones on transition and peripheral zone volumes of prostate and benign prostatic hyperplasia in twins. J Clin Endocrinol Metab. 82: Meikle AW, Stephenson RA, Lewis CM, Wiebke GA, Middleton RG Age, genetic, and nongenetic factors influencing variation in serum sex steroids and zonal volumes of the prostate and benign prostatic hyperplasia in twins. Prostate. 33: Meikle AW, Stephenson RA, McWhorter WP, Skolnick MH, Middleton RG Effects of age, sex steroids, and family relationships on volumes of prostate zones in men with and without prostate cancer. Prostate. 26: Roehrborn CG, Boyle P, Gould AL, Waldstreicher J Serum prostatespecific antigen as a predictor of prostate volume in men with benign prostatic hyperplasia. Urology. 53: Roehrborn CG, Oesterling JE, Olson PJ, Padley RJ Serial prostatespecific antigen measurements in men with clinically benign prostatic hyperplasia during a 12-month placebo-controlled study with terazosin. HYCAT Investigator Group. Hytrin Community Assessment Trial. Urology. 50: Mettlin C, Chesley AE, Murphy GP, et al Association of free PSA percent, total PSA, age, and gland volume in the detection of prostate cancer. Prostate. 39: Di Silverio F, D Eramo G, Buscarini M, et al The role of prostate specific antigen and its derivatives (age-specific PSA, PSA density, velocity, free/total PSA) in the diagnosis of prostate cancer. Minerva Urol Nefrol. 50: Maeda H, Arai Y, Okubo K, Aoki Y, Okada T, Maekawa S Value of the free to total prostate specific antigen ratio and prostate specific antigen density for detecting prostate cancer in Japanese patients. Int J Urol. 5: Sakamoto W, Iwata H, Kamikawa S, et al Role of the transition zone for elevating serum prostate-specific antigen in benign prostatic hyperplasia. Int J Urol. 5: Tarcan T, Ozveri H, Biren T, Turkeri L, Akdas A Evaluation of prostate specific antigen density and transrectal ultrasonography-guided biopsies in 100 consecutive patients with a negative digital rectal examination and intermediate serum prostate specific antigen levels. Int J Urol. 4: Diaz F, Morell M, Rojo G Usefulness of PSA density and PSA excess in the differential diagnosis between prostate cancer and benign prostatic hypertrophy. Int J Biol Markers. 11: Thon WF, Gadban F, Truss MC, Kuczyk M, Hartmann U, Jonas U Prostate-specific antigen density a reliable parameter for the detection of prostate cancer? World J Urol. 14: Filella X, Alcover J, Molina R, Carrere W, Carretero P, Ballesta AM Usefulness of prostate-specific antigen density as a diagnostic test of prostate cancer. Tumor Biol. 17: Uno H Clinical significance of PSA-density in differential diagnosis between BPH and early stages prostate cancer. Nippon Hinyokika Gakkai Zasshi. 86: Tiranti D, Annoscia S, Montefiore F, Boccafoschi C Usefulness of PSA density (PSAD) in the differential diagnosis between prostatic adenocarcinoma and benign prostatic hypertrophy. Arch Ital Urol Androl. 66: Carter HB, Pearson JD, Waclawiw Z, et al Prostate-specific antigen variability in men without prostate cancer: effect of sampling interval on prostate-specific antigen velocity. Urology. 45: Carter HB, Pearson JD PSA velocity for the diagnosis of early prostate cancer. A new concept. Urol Clin North Am. 20:

5 1276 BANSAL ET AL. JCE&M 2000 Vol 85 No Brawer MK, Meyer GE, Letran JL, et al Measurement of complexed PSA improves specificity for early detection of prostate cancer. Urology. 52: Lin DW, Noteboom JL, Blumenstein BA, Ellis WJ, Lange PH, Vessella RL Serum percent free prostate-specific antigen in metastatic prostate cancer. Urology. 52: Partin AW, Carter HB The use of prostate-specific antigen and free/total prostate-specific antigen in the diagnosis of localized prostate cancer. Urol Clin North Am. 23: Oesterling JE, Jacobsen SJ, Klee GG, et al Free, complexed and total serum prostate specific antigen: the establishment of appropriate reference ranges for their concentrations and ratios. J Urol. 154: Catalona WJ, Partin AW, Slawin KM, et al Use of the percentage of free prostate-specific antigen to enhance differentiation of prostate cancer from benign prostatic disease: a prospective multicenter clinical trial. J Am Med Assoc. 279: Partin AW, Catalona WJ, Southwick PC, Subong EN, Gasior GH, Chan DW Analysis of percent free prostate-specific antigen (PSA) for prostate cancer detection: influence of total PSA, prostate volume, and age. Urology. 48: McWhorter WP, Hernandez AD, Meikle AW, et al A screening study of prostate cancer in high risk families. J Urol. 148: Meikle AW, Stanish WM, Taylor N, Edwards CQ, Bishop CT Familial effects on plasma sex-steroid content in man: testosterone, estradiol and sex hormone-binding-globulin. Metabolism. 31: Meikle AW, Bishop DT, Stringham JD, West DW Quantitating genetic and nongenetic factors that determine plasma sex-steroid variation in normal male twins. Metabolism. 35: Wu JT, Wilson L, Zhang P, Meikle AW, Stephenson R Correlation of serum concentrations of PSA-ACT complex with total PSA in random and serial specimens from patients with BPH and prostate cancer. J Clin Lab Anal. 9: Falconer D Introduction to quantitative genetics, 2nd ed. London: Longman. 38. Page WF, Braun MM, Partin AW, Caporaso N, Walsh P Heredity and prostate cancer: a study of World War II veteran twins. Prostate. 33: