What's known on the subject? and What does the study add?
As most urologist known, obesity significantly lowers serum PSA levels. So there is some concern about delayed diagnosis of prostate cancer in obese men.
In the present study, we found that the accuracy level of PSA for detecting prostate cancer was not significantly different between different obesity levels. A well-designed study adjusting for several factors, e.g. diet, exercise, medication and comorbidity, which may possibly compensate for the associated effects on PSA levels, is needed for confirmation of the present findings.
To investigate prostate-specific antigen (PSA) accuracy in detecting prostate cancer according to body mass index (BMI) in Asian men with a PSA level of <30 ng/mL using contemporary multicore (≥12) prostate biopsy.
Patients and Methods
We reviewed the records of 3471 patients, whose initial PSA levels were <30 ng/mL, who underwent multicore (≥12) transrectal ultrasound-guided prostate biopsy between January 2004 and May 2011.
BMI was categorised as performed previously for the Asian population: <23, 23–24.9, 25–29.9, and ≥30 kg/m2. PSA accuracy for detecting prostate cancer in each BMI group was assessed based on the receiver operating characteristics-derived area under the curve.
The mean age and median PSA level were inversely associated with BMI; the median PSA level in each BMI category was 7.84, 7.75, 7.33 and 5.79 ng/mL, respectively (P < 0.001).
In all, prostate cancer was detected from biopsy in 1102 (31.7%) patients.
The PSA accuracy for predicting prostate cancer in all patients was estimated to be 0.607, and PSA accuracies in each BMI category were 0.638, 0.572, 0.613 and 0.544, respectively; there was no significant difference among the groups in terms of PSA accuracy.
The accuracy of PSA in predicting prostate cancer did not change regardless of BMI category in Asian men.
However, as patients with higher BMIs had lower PSA levels than those with lower BMIs, it can therefore be suggested that the PSA threshold should be lower in obese men to discriminate between prostate cancer and benign conditions in the real clinical situation.
Race, family history and age are the established epidemiological risk factors associated with the development of prostate cancer . However, the large geographical variation in prostate cancer risk suggests that lifestyle factors may play a significant role in the pathogenesis of cancer, particularly prostate cancer [2, 3].
Obesity, which is rapidly becoming prevalent around the world, has been linked to the development of various cancers [4, 5]; however, controversies continue about the potential association between obesity and prostate cancer [6-12]. In Western series, some have reported that obesity is associated with a higher rate of prostate cancer detection via prostate biopsy, while others have found no such association. When considering the fact that obese men have been reported to have lower serum PSA levels relative to normal-weight men in population-based studies, the idea of screening cohorts becomes even more complicated. Furthermore, the prevalence of obesity may well be different for Asian men compared with those of other races. Compared with their Western counterparts, fewer Asian men can be categorised as obese according to the widely accepted body mass index (BMI)-based definition (≥30 kg/m2) .
To date, there have been some studies that have investigated the effects of BMI on prostate cancer detection, but few studies have evaluated PSA accuracy for prostate cancer detection according to BMI, especially in Asian men. So we investigated PSA accuracy for detecting prostate cancer according to the severity of obesity among Korean men with PSA levels of <30 ng/mL who underwent prostate biopsy using the contemporary multicore (≥12) approach.
Patients and Methods
After obtaining Institutional Review Board approval, we reviewed the records of men who underwent multicore (≥12) TRUS-guided biopsy of their prostates at our institution between January 2004 and May 2011. All men underwent TRUS prostate biopsies becuase of elevated PSA levels (≥3 ng/mL), abnormal DREs or hypoechoic lesions, as detected using TRUS. In all men, the prostate was routinely biopsied bilaterally near the base, mid-gland region, and apex, with at least six biopsies per side. If necessary, additional biopsies were obtained to evaluate suspicious lesions.
Of 3893 men whose initial PSA levels were <30 ng/mL enrolled in our database of TRUS-guided biopsy, men who had undergone prior biopsies at other institutions (354 men) or the surgical treatment of prostate disease before receiving a biopsy at our institution (48) were excluded from the study, as were those for whom relevant data were missing (20). For patients who had a prostate biopsy more than once at our institution, only data from the initial biopsy were analysed. Accordingly, a total of 3471 men were included in the present study.
BMI was categorised as performed previously for the Asian population: <23, 23–24.9, 25–29.9 and ≥30 kg/m2 . According to the BMI categories, the accuracy of PSA for detecting prostate cancer in each group was assessed based on the receiver operating characteristics (ROC)-derived area under the curve (AUC). The statistical significance of the differences between various predictive accuracy estimates was compared via Mantel-Haenszel tests. Continuous variables, e.g. age, PSA level and prostate volume were analysed using simple linear regression analysis with BMI as a continuous variable and categorical variables, e.g. number of abnormal DREs (abnormal vs normal), number of hypoechoic TRUS findings and number of prostate cancer detected from biopsy, were analysed using linear-by-linear association (P for trend) in assessing differences according to the BMI groups. Multivariate logistic regression analysis was used to examine the association between BMI and prostate cancer as determined by TRUS biopsy, after adjusting for age, prostate volume, PSA level, and DRE findings. The PSA level and prostate volume were analysed after logarithmic transformations in multivariate analysis. A two-tailed P < 0.05 was considered to indicate statistical significance in all analyses.
The patients' characteristics are shown in Table 1. For the 3471 men analysed in the present study, the mean age was 64.1 years, the mean BMI was 24.3 kg/m2, the mean prostate volume was 46.1 mL, and the median PSA level was 6.0 ng/mL. The number of patients with abnormal DRE findings was 536 (25.6%) among 2096 men for whom sufficient data were available. The number of patients with hypoechoic TRUS findings was 626 (21.6%) of 2864. In all, prostate cancer was detected from the biopsy in 1102 (31.7%) men, and high-grade (Gleason score ≥ 7) prostate cancer was detected in 538 (15.5%) men.
Table 1. The patients' characteristics.
Mean (range) age, years
Mean (range) BMI, kg/m2
Median (range) PSA level, ng/mL
Abnormal DRE, n/N (%)
Hypoechoic TRUS finding, n/N (%)
Mean (range) prostate volume, mL
No. biopsy cores obtained, n (%)
Prostate cancer detected from biopsy, n (%)
High-grade (Gleason score ≥ 7) disease detected from biopsy, n (%)
When applying the aforementioned BMI categories, the numbers of patients with BMI of <23, 23–24.9, 25–29.9, and ≥30 kg/m2 were 988 (28.5%), 1126 (32.4%), 1299 (37.4%), and 58 (1.7%), respectively (Table 2). The mean age and median PSA level were inversely associated with BMI; the median PSA level in each BMI category was 7.84, 7.75, 7.33 and 5.79 ng/mL, respectively (P < 0.001). Prostate volume was positively associated with BMI category (43.6, 44.7, 48.8 and 49.2 mL, respectively, P < 0.001). The percentage of men in each BMI category in whom prostate cancer was detected was not different: 336 (34.0%) in the BMI < 23 kg/m2 category, 348 (31.2%) in the BMI 23–24.9 kg/m2 category, 403 (31.0%) in the BMI 25–29.9 kg/m2 category and 15 (25.9%) in the BMI ≥ 30 kg/m2 category (P for trend 0.298). However, as shown in Table 3, BMI was one of the significant factors predicting prostate cancer after adjustments for age, PSA level, prostate volume, DRE abnormality and hypoechoic lesions as detected by TRUS in multivariate analysis (odds ratio [OR] 1.054, 95% CI 1.004–1.106, P = 0.034).
Table 2. The patients' characteristics according to BMI.
BMI category, kg/m2
*P value by simple linear regression analysis. †P for trend by linear-by-linear association analysis.
Table 3. Multivariate analysis of factors predicting prostate cancer in men undergoing multi-core (≥12) prostate biopsy.
PSA level, ng/mL
DRE finding (normal vs abnormal)
Hypoechoic lesion in TRUS
Prostate volume, mL
Table 4 shows the accuracy of PSA level in predicting prostate cancer from biopsy samples. Among all patients, the accuracy of PSA in predicting prostate cancer, as estimated by the accumulation of AUC of the ROC, was 60.7%. PSA accuracies in each BMI category were 63.8%, 57.2%, 61.3% and 54.4%, respectively (Fig. 1). All comparisons of PSA accuracies did not differ significantly between each BMI category (Table 4).
Table 4. The accuracy of PSA level in predicting prostate cancer according to BMI and comparison of each predictive accuracy.
PSA accuracy for the detection of prostate cancer according to BMI
ROC area (se)
Comparison of PSA accuracy in predicting prostate cancer of each BMI category
<23 vs 23–24.9
23–24.9 vs 25–29.9
25–29.9 vs ≥30
In the present study, PSA level, patient age and prostate volume were significantly associated with BMI. Although a higher BMI was significantly associated with an increased risk of prostate cancer detection when adjusted for age, PSA level, DRE findings, and prostate volume, the accuracy of PSA in predicting prostate cancer did not change significantly according to BMI among a contemporary cohort of Korean men undergoing prostate biopsy using a contemporary multicore (≥12) approach. There have been some studies that have investigated the effects of BMI on prostate cancer detection from Western and Asian areas [1, 6, 7, 11], but few studies have evaluated PSA accuracy for prostate cancer detection according to BMI.
The inverse relationship between BMI and serum PSA levels has been established [15, 16]. Investigators have proposed that this effect may be due to decreased testosterone levels in obese men [17, 18]. However, recently, researchers have proposed an alternative hypothesis that the inverse relationship between BMI and serum PSA level could be explained by haemodilution [19, 20]. Banez et al.  reported that the PSA concentration decreased significantly with BMI, while the estimated total PSA mass did not, suggesting that the lower PSA concentrations in overweight and obese men could be because men with greater BMIs also tend to have larger plasma volumes when compared with other patients with prostate cancer who have undergone radical prostatectomy. Rundle and Neugut  confirmed the findings of Banez et al.  in a cohort of healthy employees undergoing annual physical examinations. The results showed that estimated plasma volumes could be used to accurately predict mean PSA concentrations in obese and morbidly obese men. The haemodilution theory has been further validated in baseline data from 28 380 men enrolled in the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial . The PSA concentration decreased significantly with increasing BMI (P < 0.001); however, plasma volume also increased with increasing BMI, and PSA mass showed no association with BMI (P = 0.10) . Therefore, low PSA levels in obese men might decrease the sensitivity of prostate cancer screening, which would lead to a delayed diagnosis with an unfavourable prognosis [4, 22, 23]. In the present study, the PSA level was also decreased inversely with BMI, but PSA accuracy in discriminating men with prostate cancer among those who were already indicative for prostate biopsy in the real clinical situation was not decreased significantly with BMI. Conversely, BMI was a significant predictor for prostate cancer in multivariate analysis, which might be due to a time effect produced by haemodilution reducing serum PSA levels, so that obese men with prostate cancer slowly reached the PSA threshold level indicative of prostate biopsy.
Rundle and Neugut  proposed that alternative PSA thresholds should be used for prostate cancer screening in obese men. According to their report, a 4.536-kg weight gain causes a −0.028 ng/mL change in PSA level. This is closely matched by the Prostate Cancer Prevention Trial results, in which a 4.536-kg weight gain caused a –0.024 ng/mL change in PSA level. Hekal and Ibrahiem  also proposed a new formula to explain the obesity and PSA relationship, and to increase the sensitivity of PSA in detecting prostate cancer. In the present study, our investigation examined those patients for whom prostate biopsy was indicated by a PSA elevation of >3.0 ng/mL or abnormal DRE or TRUS findings in the real clinical situation; accuracy of PSA in predicting prostate cancer did not change according to BMI category. Therefore, it could be suggested that the PSA level threshold should be lower in obese men to discriminate between prostate cancer and benign conditions in the real clinical situation because PSA levels of higher BMI groups was lower than that of lower BMI groups, while the overall PSA accuracy for predicting prostate cancer did not change regardless of BMI category.
The present results show that prostate volume increases along with BMI. Most reports about the relationship between prostate volume and PSA level have reported a strong positive association. After adjusting for confounding factors, high prostate volume decreased the rate of prostate cancer detection due to the quality of the needle-core biopsy, even though the samples were obtained by multicore biopsy in the present study. Put simply, increased BMI may have a negative influence on cancer detection among patients for whom biopsy has been indicated. However, in the situation that the serum PSA level has reached a certain level indicative of biopsy, obese men might have more PSA mass as a result of the haemodilution effect. Also the aforementioned time effects of BMI for delayed detection of prostate cancer may have a positive influence on the growth of cancer volume. Therefore these two conflicting effects in obese men might offset the PSA accuracy for detecting prostate cancer, resulting in the same level of accuracy regardless of severity of obesity. The finding that the PSA accuracy level among obese patients was not lower than that among non-obese patients might mean that the PSA threshold level should be decreased because the PSA level of the obese men was lower than that of the non-obese men
There are some limitations to the present study that should be considered when interpreting the results. Previous Western studies on the impact of obesity on the detection of prostate cancer via prostate biopsy have frequently defined obesity as a BMI of ≥30 kg/m2 [25-28]. As seen from the present series, this definition is not valid when applied to Asian men. Among adult Korean men in general, <2% have been reported to have a BMI of ≥30 kg/m2. This percentage is much lower than that reported in Europe and the USA . In the present series, only 1.7% of men had a BMI of ≥30 kg/m2, a threshold that bore no significance for the cancer detection rate via biopsy among the present cohort. However, when BMI was categorised into three groups: <23 kg/m2 (988, 28.5%), 23– 24.9 kg/m2 (1126, 32.4%) and ≥25 kg/m2 (1357, 39.1%) in the present study, the results for PSA accuracy did not differ (data not shown). Nevertheless further study is needed in Asian cohorts that include more men who have a BMI of ≥30 kg/m2. There were some differences between the present study and previous studies of Western cohorts for the prostate cancer detection rate and PSA accuracy for prostate cancer detection, which may be due to racial differences. Another limitation of the present study was the failure to adjust for confounding factors, e.g. diet, exercise, medication and comorbidities, to possibly compensate for the associated effects on PSA levels. The present study is also limited by its retrospective nature. Although BMI is currently considered a convenient proxy measure of body fat, its limitations have been recognised, given its inability to distinguish fat mass from lean mass. For example, if lean body mass declines because of chronic disease or physical inactivity, adiposity may increase without a change in BMI [29, 30]. Other parameters, e.g. waist circumference or waist-to-hip ratio, may be a better indicator for obesity in a relatively mature population. However, the obvious difference in the prevalence of obesity between Asian and Western populations is well-represented in the BMI profiles of the present patients. The multicore (≥12) prostate biopsy was performed in a uniform and contemporary fashion in all of the patients.
In conclusion, the present results showed that a higher BMI significantly decreased PSA levels among Korean men undergoing contemporary multicore prostate biopsy. The accuracy of PSA in predicting prostate cancer did not change regardless of BMI category. However, as patients with higher BMI had lower PSA levels than those with lower BMIs, it can therefore be suggested that the PSA threshold should be lower in obese men to discriminate between prostate cancer and benign conditions in the real clinical situation. Further studies are needed to determine the adjusted PSA threshold level for obese men among a large cohort evenly distributed in BMI.
This study has been supported by a grant from Seoul National University Bundang Hospital Research Fund.
Conflict of Interest
None of the authors have any conflicts of interest with any institution or product.