Background: A hematocrit (HCT)-based and a body surface area (BSA)–based equations were applied for plasma volume (PV) estimation, respectively, to confirm and quantify the hemodilution effect in prostate-specific antigen (PSA) screening among obese men. The agreement between the equations was additionally investigated.

Methods: A total of 1,444 men were retrospectively collected, with ages 40 to 65 years, PSA 0 to 4 ng/mL, and no prostate cancer. PSA mass was calculated as PSA concentration multiplied by PV. Multivariable linear regression models, theoretical models, and the Bland–Altman method were used.

Results: PSA concentration significantly decreased with increasing body mass index (BMI; β = −0.011, P < 0.001); however, PSA mass estimated by HCT- (β = 0.004, P = 0.132) and BSA (β = −0.003, P = 0.094)-based equations remained consistent. A screening PSA of 4.0 ng/mL in nonobese men was found to be corresponding to 3.32 and 3.68 ng/mL in obese men extrapolated by PV on the basis of HCT and BSA, respectively. Moreover, the mean (95% confidence interval) difference of PV between the two equations was 0.33 (−0.06 to 0.73) L.

Conclusions: The inverse relationship between PSA concentration and BMI might be explained by a hemodilution effect among obese men. There is significant variation in PV calculated by the two equations.

Impact: A value between 3.32 and 3.68 ng/mL might be recommended for PSA screening in middle-aged obese Asian men. Cancer Epidemiol Biomarkers Prev; 21(10); 1731–4. ©2012 AACR.

It has been reported (1–3) that lower serum prostate-specific antigen (PSA) levels among obese men suggest that obesity might have a negative effect on screening for prostate cancer by lowering serum PSA levels of these men. Recent studies (3–5) showed that although PSA concentration decreased significantly with the increase of body mass index (BMI), the estimated PSA mass did not, indicating that the lower PSA concentration observed in obese men could be merely because of the fact that men with greater BMI also tended to have larger plasma volume (PV). In these studies, estimated PV and PSA mass were based on a body surface area (BSA) equation. However, a recent study indicated that there are significant variation in the PV calculated by different BSA-based equations, which are further magnified in the sequential calculation of PSA mass (6). Here, another equation for PV estimation was proposed, which has not been chosen in similar studies before but widely applied for drug administration, with a new introduced variable hematocrit (HCT%; refs. 6, 7):

  1. PV (L) = 0.07 × weight (kg) × (1 − HCT)

  2. PSA mass (μg) = PSA concentration × PV

Meanwhile, a commonly chosen BSA-based formula (3–5) appended later was applied for method comparison:

  1. BSA (m2) = body weight (kg)0.425 × height (m)0.725 × 0.2025

  2. PV (L) = BSA (m2) × 1.670

  3. PSA mass (μg) = PSA concentration × PV

We are going to investigate:

  1. The evidence of hemodilution effect among obese men and the implications for PSA screening;

  2. The agreement between the estimates by using the HCT-based (7, 8) and the BSA-based (3–5) equations.

Participants selection

This was a retrospective study of participants presenting at our unit for health management from May to November 2008. A total of 4,312 consecutive males, ages 40 to 65 years, underwent a physical examination during the initial consultation. The information obtained was then prospectively collected in a computer database. After 3 years (May 2009 to November 2011), follow-ups were done once in a year for this cohort. We excluded participants with missing data due to rejection of uncomfortable examinations, including digital rectal examination and blood collection, or unselection of the items by self-funded participants, or having part of the laboratory examinations detected recently in another clinical center with unshared results. The missing data were detailed as follows: missing BMI (75 cases), absence of HCT (295 cases) or serum PSA concentration (1,037 cases), no recorded history of benign prostate problems (1,174 cases), no data of alanine aminotransferase (ALT; 342 cases), aspartate aminotransferase (AST; 342 cases), or serum creatinine (Cr; 183 cases). There were overlaps among these missing data. People with one of the following criteria were additionally excluded:

  1. PSA concentration higher than 4 ng/mL (92 cases);

  2. Confirmed prostate cancer previously or within the 3-year follow-up (5 cases);

  3. The use of medication affecting PSA concentration (82 cases);

  4. A previous surgical procedure related to benign prostate problem (30 cases).

The remaining 1,444 males without clinically detectable prostate cancer were included in the present analysis. Our local Institutional Review Board approved the survey before it was conducted.

Clinical variables

Standing height and weight were measured with emptied pockets and without shoes; the BMI was calculated as usual. Subjects were divided into 2 BMI categories of 27.4 or less and ≥27.5 or more according to the new Asian classification (9). The HCT was measured on a Hematology Analyzer (Beckman COULTER GEN'S System 2); ALT, AST, and serum Cr on a Hitachi 7600 modular chemistry analyzer; serum PSA concentration on a Beckman Coulter DXI 800 Immunoassay Analyzer. All the above assessments were measured by qualified staffs at our health management center. PVs and PSA masses were estimated by using the earlier described equations (3–5, 7, 8).

Statistical analyses

Statistical analyses were conducted by using the SPSS 13.0 software. PSA and PSA mass were logarithmically transformed to provide a more normal data distribution. Multivariable linear regression models were conducted to investigate the changes of PV, PSA concentration, and masses estimated by respective equations with the increasing of BMI. Age was adjusted in the model when PV was defined as a dependent variable. Benign prostate problems, AST, ALT, and Cr were additionally adjusted with PSA or mass as dependent variable, as it was shown that these independent variables might contribute to the changes of PSA concentration or mass in previous literatures (3, 10).

A mathematical model was established for the calculation of a serum PSA level in an obese man that corresponds to a PSA of 4.0 ng/mL in a nonobese man (5). The geometric means of height, weight, and HCT were obtained for the nonobese and obese group, respectively, to calculate the PV of each group (PVNonobese and PVObese). As PSA mass was assumed to be consistent according to BMI group in healthy cohorts (3, 5), an equation can be extrapolated: the cutoff value of PSA concentration predicted for an obese man in this cohort = 4.0 ng/mL × PVNonobese/PVObese. The HCT-based and BSA-based PVs were put into the model, respectively.

To analyze the agreement between the 2 equations for PV estimation, Bland and Altman method was additionally conducted (11).

Figure 1 shows a distribution of men by BMI and PSA for the whole cohort. Table 1 shows the relationships among PSA concentration, PV, PSA mass, and BMI. Approximately 17.7% of males were classified as obese. When BMI was analyzed as a continuous variable, logarithmically transformed PSA levels were inversely associated with BMI significantly (β = −0.011, P < 0.001). Means of estimated PVs by using HCT- and BSA-based equations, increased with increasing BMI category, from 2.65 L in nonobese men to 3.18 L in obese men, and from 3.03 to 3.28 L, respectively (P < 0.05). The sequentially estimated PSA masses, however, did not show a significant association with BMI with the adjustment of age, benign prostate problem, ALT, AST, and Cr (β = 0.004, P = 0.132 for HCT-based equation; β = −0.003, P = 0.094 for BSA based one).

Figure 1.

Distribution of men by BMI and PSA for this cohort.

Figure 1.

Distribution of men by BMI and PSA for this cohort.

Close modal
Table 1.

PSA concentration, HCT-based, and BSA-based PVs, and PSA masses according to BMI category

WHO Asian BMI category (kg/m2)
≤27.4 (Nonobese group)≥27.5 (Obese group)P for trend
No men (%) 1,189 (82.3) 255 (17.7) N/A 
PSA, geometric mean (IQR; ng/mL)a 0.87 (0.59–1.26) 0.73 (0.47–1.09) <0.001 
Plasma volume, mean (SD; L)b    
HCT-based 2.65 (0.31) 3.18 (0.30) <0.001 
BSA-based 3.03 (0.20) 3.28 (0.19) <0.001 
PSA mass, geometric mean (IQR; μg)a    
HCT-based 2.28 (1.52–3.34) 2.31 (1.51–3.50) 0.132 
BSA-based 2.62 (1.77–3.81) 2.40 (1.56–3.63) 0.094 
WHO Asian BMI category (kg/m2)
≤27.4 (Nonobese group)≥27.5 (Obese group)P for trend
No men (%) 1,189 (82.3) 255 (17.7) N/A 
PSA, geometric mean (IQR; ng/mL)a 0.87 (0.59–1.26) 0.73 (0.47–1.09) <0.001 
Plasma volume, mean (SD; L)b    
HCT-based 2.65 (0.31) 3.18 (0.30) <0.001 
BSA-based 3.03 (0.20) 3.28 (0.19) <0.001 
PSA mass, geometric mean (IQR; μg)a    
HCT-based 2.28 (1.52–3.34) 2.31 (1.51–3.50) 0.132 
BSA-based 2.62 (1.77–3.81) 2.40 (1.56–3.63) 0.094 

aAdjusted for age, history of prostate problems, ALT, AST, and Cr.

bAdjusted for age.

The mathematical modeling procedures by using the obtained geometric means of weight, height, and HCT, showed that a PSA concentration of 4 ng/mL in nonobese group equates to a PSA masses of 10.52 μg (4 ng/mL × 2.63 L) by using HCT-based PV and 12.08 μg (4 ng/mL × 3.02 L) by BSA-based equation. Respective PVs of 3.19 and 3.28 L for obese group were also obtained according to the methods mentioned earlier, resulting in 2 cutoff PSA concentrations for obese men, 3.32 and 3.68 ng/mL.

Bland and Altman plot was used to examine how the means of PV between the 2 formulae changed relative to their difference (Fig. 2). The BSA-based formula produced higher bias at a low PV (overestimation) than that from HCT-based equation [mean (95% confidence interval) difference: 0.33 (−0.06 to 0.73) L; r = −0.542, P < 0.001). In fig. 2, there were 67 data points (bias) beyond the ±1.96 SD limits. The maximum difference between the 2 equations was 0.89 L.

Figure 2.

Bland and Altman plot of the difference in PV as a function of the mean PV on the basis of the BSA and HCT formulae.

Figure 2.

Bland and Altman plot of the difference in PV as a function of the mean PV on the basis of the BSA and HCT formulae.

Close modal

We have confirmed in this Chinese cohort the widely reported inverse association between PSA levels and BMI (12). Specifically, 2 equations, based on HCT and BSA, respectively, are selected for PV and PSA mass estimation. The results obtained by both equations indicated that although PSA concentration decreased with the increasing BMI, however, PSA mass remained consistent. This suggests that hemodilution from larger PV may be responsible for the lower PSA values observed among obese men. Also, in 2 mathematical models by using HCT- and BSA-based equations for PV estimation, serum PSA concentrations of 3.32 and 3.68 ng/mL in obese men that correspond to a screening PSA of 4.0 ng/mL in nonobese men were predicted, respectively. The latter was consistent with the study done by Rundle and Neugut' (3.65 ng/mL; ref. 5) by using the same BSA-based equation, whereas the former one seemed to be lower. That might be because of the overestimation of the dilution power of PV estimated by the HCT-based equation and underestimation of that calculated by BSA-based formula when compared with mean PV. The “real” cutoff value for obese men might just lie between the 2 values.

For further discussion on the problem, the agreement between the HCT- and BSA-based equations for PV estimation was investigated. The Bland and Altman plot displays considerable lack of agreement between the 2 equations, with discrepancies of PV up to 890 mL, which will undoubtedly be greatly magnified in the sequential calculation of PSA mass. Because of the absence of PV determined by “Golden criteria,” we could not tell which equation is more suitable for clinical application. However, the discrepancies call for greater attention to further study on the problem.

One alternative explanation for the decreased PSA concentration seen in the obese men may be a decrease in circulating levels of androgens that are more likely to be found in obese men (13). However, previous studies have shown that only for hypogonadal men, who underwent testosterone replacement treatment, PSA levels increases (14). Although for healthy young men suffering testosterone administration, PSA did not change (15). The earlier data might indicate that the influence of testosterone on PSA level is not consistent and maybe dependent on the gonad function. Moreover, given that PSA is under androgenic control, obese men would be expected to have lower PSA mass. However, several studies, including ours, confirm that obese men had similar or even higher PSA mass in cohorts of both healthy males and patients with prostate cancer (3, 4). On the basis of these studies, we proposed that the PSA level might be mainly controlled by both hormones and hemodilution, which might be useful for explaining the conflicting results on the relationship between BMI and PSA found in other studies (16, 17). Our data support that in the present cohort, hemodilution predominates and that hormonal effects are rendered negligible. Unfortunately, we do not have testosterone levels in all of the men in the cohort, and further studies are still needed to test the effects of hormones on PSA level.

The inverse relationship between PSA concentration and BMI might be explained by a hemodilution effect among obese men. There is significant variation in the PV calculated by the 2 equations.

No potential conflicts of interest were disclosed.

Conception and design: F. Li

Development of methodology: F. Li

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): F. Li, Z. Shen, Y. Lu, J. Yun, Y. Fan, C. Song

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): F. Li

Writing, review, and/or revision of the manuscript: F. Li

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): F. Li

Study supervision: F. Li

The authors greatly appreciate the technical assistance provided by Dr. Weizhu Shi, Lanping Li and Weiqi Zhang.

1.
Werny
DM
,
Thompson
T
,
Saraiya
M
,
Freedman
D
,
Kottiri
BJ
,
German
RR
, et al
Obesity is negatively associated with prostate-specific antigen in U.S. men, 2001–2004
.
Cancer Epidemiol Biomarkers Prev
2007
;
16
:
70
6
.
2.
Fowke
JH
,
Signorello
LB
,
Chang
SS
,
Matthews
CE
,
Buchowski
MS
,
Cookson
MS
, et al
Effects of obesity and height on PSA and percent free PSA levels among African-American and Caucasian men
.
Cancer
2006
;
107
:
2361
67
.
3.
Grubb
RL
 III
,
Black
A
,
Izmirlian
G
,
Hickey
TP
,
Pinsky
PF
,
Mabie
JE
, et al
Serum prostate-specific antigen hemodilution among obese men undergoing screening in the prostate, lung, colorectal, and ovarian cancer screening trial
.
Cancer Epidemiol Biomarkers Prev
2009
;
18
:
748
51
.
4.
Bañez
LL
,
Hamilton
RJ
,
Partin
AW
,
Vollmer
RT
,
Sun
L
,
Rodriguez
C
, et al
Obesity-related plasma hemodilution and PSA concentration among men with prostate cancer
.
JAMA
2007
;
298
:
2275
80
.
5.
Rundle
A
,
Neugut
AI
. 
Obesity and screening PSA levels among men undergoing an annual physical exam
.
Prostate
2008
;
68
:
373
80
.
6.
Sivanandam
A
,
Siva
S
,
Bhandari
M
,
Menon
M
. 
Variance inflation in sequential calculations of body surface area, plasma volume, and prostate-specific antigen mass
.
BJU Int
2008
;
102
:
1573
80
.
7.
Kaplan
AA
. 
A simple and accurate method for prescribing plasma exchange
.
ASAIO Trans
1990
;
36
:
M597
9
.
8.
Haskell
MJ
,
Islam
MA
,
Handelman
GJ
,
Peerson
JM
,
Jones
AD
,
Wahed
MA
, et al
Plasma kinetics of an oral dose of [2H4]retinyl acetate in human subjects with estimated low or high total body stores of vitamin A
.
Am J Clin Nutr
1998
;
68
:
90
5
.
9.
WHO Expert Consultation
. 
Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies
.
Lancet
2004
;
363
:
157
63
.
10.
Ahn
JO
,
Ku
JH
. 
Relationship between serum prostate-specific antigen levels and body mass index in healthy younger men
.
Urology
2006
;
68
:
570
4
.
11.
Bland
JM
,
Altman
DG
. 
Statistical methods for assessing agreement between two methods of clinical measurement
.
Lancet
1986
;
8
:
307
10
.
12.
Baillargeon
J
,
Pollock
BH
,
Kristal
AR
,
Bradshaw
P
,
Hernandez
J
,
Basler
J
, et al
The association of body mass index and prostate-specific antigen in a population-based study
.
Cancer
2005
;
103
:
1092
5
.
13.
Freedland
SJ
,
Platz
EA
,
Presti
JC
,
Aronson
WJ
,
Amling
CL
,
Kane
CJ
, et al
Obesity, serum prostate specific antigen and prostate size: implications for prostate cancer detection
.
J Urol
2006
;
175
:
500
4
.
14.
Cooper
CS
,
Perry
PJ
,
Sparks
AE
,
MacIndoe
JH
,
Yates
WR
,
Williams
RD
, et al
Effect of exogenous testosterone on prostate volume, serum and semen prostate specific antigen levels in healthy young men
.
J Urol
1998
;
159
:
441
3
.
15.
Bhasin
S
,
Woodhouse
L
,
Casaburi
R
,
Singh
AB
,
Bhasin
D
,
Berman
N
, et al
Testosterone dose-response relationships in healthy young men
.
Am J Physiol Endocrinol Metab
2001
;
281
:
1172
81
.
16.
Thompson
IM
,
Leach
R
,
Troyer
D
,
Pollock
B
,
Naylor
S
,
Higgins
B
, et al
Relationship of body mass index and prostate specific antigen in a population-based study
.
Urol Oncol
2004
;
22
:
127
31
.
17.
Hutterer
G
,
Perrotte
P
,
Gallina
A
,
Walz
J
,
Jeldres
C
,
Traumann
M
, et al
Body mass index does not predict prostate-specific antigen or percent free prostate-specific antigen in men undergoing prostate cancer screening
.
Eur J Cancer
2007
;
43
:
1180
7
.