Clinicopathological presentation and ancestry
All recruited men visited a Southern African Prostate Cancer Study (SAPCS) urology clinic due to urological complaints, including erectile dysfunction, lower urinary tract symptoms, swollen scrotal region, dysuria or hematuria, and were recommended for prostate biopsy. Through histopathological analyses of 12 or more biopsy cores per patient, PCa diagnosis was confirmed via Gleason score and International Society of Urologic Pathologists (ISUP) group grading, resulting in 120 men receiving a positive diagnosis (case) and 130 a negative diagnosis (control; Table 1). Patients with PCa were further grouped according to their ISUP grading (or Gleason score) as high-risk (HRPCa, ISUP 3 - 5, n = 51) or low-risk (LRPCa, ISUP 1 - 2, n = 69) PCa. The mean age at presentation was 66.5 years for patients with PCa and 66.2 years for controls. As expected, prostate-specific antigen (PSA) levels were significantly elevated in cases versus controls (102.4 vs. 31.4 µg/L, P <0.001) and in HRPCa vs LRPCa and no PCa groups (Figure S1A). Concurring with previous data reported for Black South African men [5], PSA levels far exceeded global averages for both groups, while appear greatest for men with PSAs ≥20 µg/L (68.4% cases vs. 21.6% controls, P = 0.009; Figure S1B).
Notably, the largest racial-based study out of the United States recently showed Black men without PCa to have the highest PSA levels than all other racial groups [28]. Here, African ancestry was further defined through two generational self-identification, with the inclusion of one or more of the following southern Bantu or Black South African ethno-linguistic groupings defined via Guthrie Zone-S linguistic groups as S20-Venda (Tshivenda speakers), S30-Sotho-Tswana (Sesotho, Sepedi, Setswana), S40-Nguni (isiNdebele, isiXhosa, isiZulu, siSwazi), and S50-Tsonga (Xitsonga, including in our study ethnically reported Shangaan). Observing no significant clinicopathological, including PSA levels (Figure S1C), between the ethno-linguistic groups, with Sotho-Tswana speakers representing the largest study contributors (52.4%, 131/250; Table 1) participants were comfortably classified in downstream analyses as a single Southern Bantu or Black South African representative population.
Testosterone and lipid levels and PCa status
Controlling for time of sampling, defined in this study as prior to 10h00, men presenting at the clinic were tested for total serum testosterone, including albumin and sex hormone binding globulin (SHBG) to determine free and bioavailable testosterone, as well as patient- and time-matched serum cholesterol and triglyceride levels. Overall, 18% of the study population had low (<300 ng/dL) total testosterone, and 5.6% had high (>1,000 ng/dL) total testosterone (Table 1). Men with PCa (525.6 ± 308.4 ng/dL) did not have significantly lower testosterone levels than men without PCa (574.6 ± 286.5 ng/dL). The mean cholesterol and triglyceride levels were similar between men with and without PCa (5.1 ± 1.2 vs. 5.2 ± 1.3 mmol/L; triglyceride levels: 1.66 ± 0.81 mmol/L, respectively). The proportion of patients with high cholesterol levels (>6.18 mmol/L) did not differ significantly between PCa patients and healthy controls (15% vs. 14.6%).
Table 1. Clinical and hormone/lipid biochemical characteristics of the study cohort of 250 Black South African men, either with (positive) or without (negative) clinicopathologically confirmed prostate cancer
|
PCa positive (n = 120)
|
PCa negative
(n = 130)
|
P-value
|
Age Number
Mean years (Range)
|
119
66.5 (43 – 89)
|
128
66.2 (45 – 91)
|
0.79
|
Ethno-linguistic identifier
Venda
Sotho-Tswana
Nguni
Tsonga
Other
Unknown
PSA
Mean µg/L ± SD Number < 4 µg/L
Number 4 – 19.9 µg/L Number 20 – 99 µg/L Number ≥ 100 µg/L
|
16 (13.3%)
69 (57.5%)
20 (16.7%)
9 (7.5%)
3 (2.5%)
3 (2.5%)
102.4 ± 208.2 2 (1.67%)
36 (30.0%) 59 (49.2%) 23 (19.2%)
|
10 (7.7%)
62 (47.7%)
29 (22.3%)
22 (16.9%)
5 (3.8%)
2 (1.5%)
31.4 ± 123.6 18 (13.9%)
84 (64.6%) 24 (18.5%) 4 (3.1%)
|
<0.0001
|
PCa rating LRPCa (ISUP 1 & 2) HRPCa (ISUP 3, 4 & 5)
|
69 (57.5%) 51 (42.5%)
|
|
|
Total testosterone Mean ng/dL ± SD Number <300 ng/dL Number 300–1000 ng/dL
Number >1000 ng/dL
|
525.6 ± 308.4 24 (20%) 91 (75.8%) 5 (4.2%)
|
574.6 ± 286.5 21 (16.2%) 100 (76.9%) 9 (6.9%)
|
0.19
|
Cholesterol
Mean mmol/L ± SD Number <5.17 mmol/L Number 5.17-6.18 mmol/L Number >6.18 mmol/L
|
5.1 ± 1.2 68 (56.7%) 34 (28.3) 18 (15%)
|
5.2 ± 1.3 72 (55.4%) 39 (30%) 19 (14.6%)
|
0.58
|
Triglycerides
Mean mmol/L ± SD Number <1.7 mmol/L Number >1.7 mmol/L
|
1.66 ± 0.81 75 (62.5%) 45 (37.5%)
|
1.66 ± 0.81 82 (63.1%) 48 (36.9%)
|
0.95
|
Abbreviations: PCa: Prostate Cancer; LRPCa: low-risk PCa; HRPCa: high-risk PCa; PSA: Prostate Specific Antigen; SD: Standard Deviation
Age-associated testosterone and lipid levels, PCa status and presentation
While total (Figure 1A), bioavailable (Figure 1B), and free (Figure 1C) testosterone decreased with age as expected, this decrease became more pronounced for cases over controls after 70 years of age. Compared with younger men, men aged 65 years with PCa had significantly lower levels of bioavailable (340.42 vs. 188.40 ng/dL, P = 0.045) and free testosterone (12.54 vs. 7.37 ng/dL ng/dL, P = 0.048), while total testosterone levels did not quit reach significance (Table 2). Irrespective of age, patients presenting with high-risk (ISUP 3 – 5) PCa had significantly lower total (Figure 1D), bioavailable (Figure 1E), and free testosterone levels (Figure 1F) than the low-risk (ISUP <3) patients and the control group (P = 0.004, P = 0.0036, and P = 0.0012, respectively). The latter holding true for men 65 years (P = 0.0057 (Figure S2A), P = 0.009 (Figure S2B), and P = 0.005 (Figure S2C), respectively), with significance increasing for men 70 years of age (P = 0.0032 (Figure 1G), P = 0.00012 (Figure 1H), and P = 0.00039 (Figure 1I), respectively). Unlike testosterone, neither cholesterol nor triglyceride levels were linked to age-associated PCa status (Figure S3A and S3B, respectively) or disease presentation (Figure S3C and S3D, respectively). Additionally, no correlation was observed between total testosterone and cholesterol levels according to PCa status (Figure S4A) or age (Figure S4B).
Age-associated testosterone levels between Black South African and American men
According to the US-based National Health and Nutritional Examination Survey (NHANES), healthy Black men (n=355) have on average 1.2-fold higher serum testosterone levels than White American men (n=631), with testosterone levels appearing to decrease more rapidly for African American men [26]. As such, NHANES provides a foundation for comparative analyses for our Black South African control data (n=130). Here we observe largely an incremental change in testosterone levels between White and Black American and Black South African men, which is greatest for men 60 years of age and over (Table 3). While the fold increase is comparable for total- (1.29) and free testosterone (1.28) for Black South African versus Black American men, the greatest change was observed for bioavailable testosterone at 1.54-fold; or 1.80-fold between Black South African and White American men.
Table 2: Mean testosterone, cholesterol and triglyceride levels in younger (<65 years) and older (≥65 years) Black South African men (n=250) defined by PCa status.
|
PCa Negative
|
PCa Positive
|
Age
|
40–64 years (n = 58)
|
≥65 years (n = 72)
|
P-value
|
40–64 years (n = 52)
|
≥65 years (n = 68)
|
P-value
|
Total T (ng/dL)
|
590.45
|
561.89
|
0.578
|
585.70
|
479.55
|
0.065
|
Free T (ng/dL)
|
9.41
|
9.50
|
0.944
|
12.54
|
7.37
|
0.048
|
Bioavailable T (ng/dL)
|
245.28
|
247.26
|
0.955
|
340.42
|
188.40
|
0.045
|
Cholesterol (mmol/L)
|
5.04
|
5.24
|
0.356
|
5.10
|
5.04
|
0.759
|
Triglycerides (mmol/L)
|
1.73
|
1.61
|
0.405
|
1.76
|
1.58
|
0,2060 0.206
|
Abbreviations: PCa: Prostate Cancer; T: Testosterone
Table 3. Testosterone levels (total, free and bioavailable) between Black South African (this study, n=130) and Black- (n=355) and White (n=631) American (NHANES study) men with diagnosis of prostate cancer
Testosterone
|
Ethnicity
|
Age 40 – 59 years
|
|
Age ≥ 60 years
|
|
|
ng/dL
|
Fold
|
ng/dL
|
Fold
|
Fold
|
Total
|
White American
|
439 (402 – 478)
|
-
|
359 (326 – 395)
|
-
|
1.22
|
Black American
|
524 (465 – 592)
|
1.191
|
432 (374 – 499)
|
1.201
|
1.21
|
Black South African
|
648.70 (451.32 – 846.08)
|
1.242
|
557.46 (506.60 – 608.33)
|
1.292
|
1.16
|
Free
|
White American
|
8.80 (7.95 – 9.73)
|
-
|
5.50 (4.99 – 6.07)
|
-
|
1.6
|
Black American
|
10.16 (8.97 – 11.52)
|
1.161
|
6.84 (6.05 – 7.72)
|
1.241
|
1.49
|
Black South African
|
10.08 (8.28 – 11.88)
|
0.992
|
8.74 (7.99 – 9.49)
|
1.282
|
1.15
|
Bioavailable
|
White American
|
209.38 (189.43 – 231.43)
|
-
|
127.52 (115.49 – 140.81)
|
-
|
1.64
|
Black American
|
231.36 (203.48 – 263.06)
|
1.111
|
149.44 (131.13 – 170.31)
|
1.171
|
1.55
|
Black South African
|
263.34 (213.13 – 313.56)
|
1.142
|
229.73 (205.48 – 253.98)
|
1.542
|
1.15
|
1 Fold increased between Black Americans and White Americans.
2 Fold increase between Black South Africans and Black Americans.