Elevated prostate specific antigen and reduced 10-year survival among a cohort of Danish men consecutively referred from primary care to an urological department during 2005–2006

Figures & data

Figure 1. Patient characteristics. During June 2005 to August 2006, 1988 patients with a variety of urological symptoms were referred from primary care in the Capital Region of Copenhagen, Area North, to The Department of Urology Frederikssund Hospital, Capital Region of Copenhagen, Area North. tPSA, total Prostate Specific Antigen; cPSA, complexed Prostate Specific Antigen; %fPSA, percentage free Prostate Specific Antigen relative to tPSA. All male patients were tested for tPSA, cPSA and %fPSA concentrations in serum. Patients with previously diagnosed PCa were excluded from the dataset. ¹⁾PCa diagnoses (N = 299) were histologically confirmed by a 6–12 core biopsy or prostate resectate. ²⁾Non-PCa diagnoses (N = 209) were histologically confirmed by a 6–12 core biopsy or prostate resectate. ³⁾Eight patients with too sparse biopsy material for Gleason staging and five patients not registred for Gleason score. ⁴⁾Ninety patients were diagnosed with another malignancy than PCa: C. Testis/Vesica Seminalis (7), C. Coli/Recti (31), and C. Urethrae/Vesica Urinaria (52). ⁵⁾Other; Obs. pro c. prostatae (25), non-malignant disease of the testes and vesica seminalis (44), prostate abscess (2), urolithiasis (4), urosepsis (2), hydronephrosis (2), proteinuria (1), gynacomastia (2), pain (7), abnormal urinary cytology (8), and nothing abnormal (7).

Figure 2. Number of patients with and without PCa within different tPSA and cPSA concentration intervals. Some 299 patients were diagnosed with Prostate Cancer (PCa), and 962 patients were diagnosed as non-PCa. tPSA denotes total Prostate Specific Antigen. cPSA denotes complexed Prostate Specific Antigen. The recommended cut-off for tPSA was 4.0 μg/L, and the recommended cut-off cPSA was 3.6 μg/L.

Table 1. Median tPSA values obtained for different age intervals including all 1261 eligible patients.

Age at sampling (years)	Number of patients	Median tPSA
19–40	14	0.76
40–50	61	0.72
50–60	190	2.10
60–70	440	4.34
70–80	400	6.10
>80	156	6.39

tPSA: total Prostate-Specific Antigen.

Table 2. Overall diagnostic performance of tPSA, cPSA and %fPSA among 1261 consecutive patients referred to the Department of Urology from Primary Care.

	tPSA	cPSA	%fPSA
Cut-off	4 μg/L^a N = 1261	3.6 μg/L^a N = 1261	30%^b N = 1261	25%^b N = 1261	20%^b N = 1261	15%^b N = 1261	10%^b N = 1261
Sensitivity	98.0% (97.1–98.6)	97.7% (96.7–98.4)	92.3% (90.7–93.7)	85.3% (83.2–87.1)	76.6% (74.2–78.8)	59.2% (56.5–61.9)	38.5% (35.8–41.2)
Specificity	62.8% (60.1–65.4)	67.7% (65.0–70.2)	35.7% (33.1–38.3)	49.8% (47.0–52.5)	65.5% (62.8–68.1)	79.7% (77.4–81.9)	90.2% (88.5–91.7)
Pvpos	45.0% (42.3–47.8)	48.4% (45.7–51.2)	30.8% (28.4–33.4)	34.6% (32.0–37.2)	40.8% (38.1–43.6)	47.6% (44.8–50.3)	55.0% (52.3–57.8)
Pvneg	99.0% (98.3–99.4)	98.9% (98.2–99.4)	93.7% (92.2–94.9)	91.6% (89.9–93.0)	90.0% (88.2–91.5)	86.3% (84.3–88.1)	82.5% (80.3–84.5)
Efficiency	71.1% (68.6–73.6)	74.8% (72.3–77.1)	49.1% (46.3–51.8)	58.2% (55.5–60.9)	68.1% (65.5–70.6)	74.9% (72.4–77.2)	78.0% (75.6–80.2)
2 × 2 table	293 TP 358 FP^c	292 TP 311 FP^c	23 FN 343 TN^c	44 FN 479 TN^c	70 FN 630 TN^c	122 FN 767 TN^c	184 FN 868 TN^c
	6 FN 604 TN^c	7 FN 651 TN^c	276 TP 619 FP^c	255 TP 483 FP^c	229 TP 332 FP^c	177 TP 195 FP^c	115 TP 94 FP^c

A total of 1261 patients with PSA measurements were eligible for the study; 299 patients were diagnosed with Prostate Cancer (PCa), and 962 patients were diagnosed as non-PCa. tPSA denotes total Prostate Specific Antigen. cPSA denotes complexed Prostate Specific Antigen. %fPSA denotes percentage free Prostate Specific Antigen relative to tPSA. The recommended cut-off for tPSA was 4.0 μg/L. The recommended cut-off cPSA was 3.6 μg/L. The cut-off for %fPSA was set to 30, 25, 20, 15 and 10%, respectively. PVpos: Predictive Value of a positive test; PVneg: Predictive Value of a negative test; TP: True Positive patients; FN: False Negative patients; FP: False Positive patients; TN: True Negative patients. Sensitivity was defined as TP/(TP + FN). Specificity was defined as TN/(TN + FP). PVpos was defined as TP/(TP + FP). PVneg was defined as TN/(TN + FN). Efficiency was defined as (TP + TN)/TP + FN + TN + FP.

^aConcentrations > the specified cut-off were classified as TP (PCa) and FP (non-PCa) patients, respectively. Concentrations ≤ the specified cut-off were classified as FN (PCa) and TN (non-PCa) patients, respectively.

^bConcentrations ≥ the specified cut-off were classified as FN (PCa) and TN (non-PCa) patients, respectively. Concentrations < the specified cut-off were classified as TP (PCa) and FP (non-PCa) patients, respectively.

^cTN and FP results were not in all cases verified by biopsy. 95% Confidence intervals are shown in italics.

Table 3. Diagnostic performance of %fPSA within different tPSA concentration intervals among 1261 consecutive patients referred to the Department of Urology from Primary Care.

tPSA concentration intervals	≤ 4 μg/L tPSA N = 610					> 4 – ≤ 10 μg/L tPSA N = 372
Cut-off %fPSA	30%^a	25%^a	20%^a	15%^a	10%^a	30%^a	25%^a	20%^a	15%^a	10%^a
Sensitivity	66.7% (62.8–70.3)	66.7% (62.8–70.3)	50.0% (46.0–54.0)	16.7% (13.9–19.8)	16.7% (13.9–19.8)	99.0% (97.4–99.6)	89.0% (85.4–91.8)	79.0% (74.5–82.9)	62.0% (57.0–66.8)	41.0% (36.1–46.1)
Specificity	45.7% (41.8–49.7)	59.9% (56.0–63.7)	75.7% (72.1–78.9)	86.9%(84.0–89.4)	94.0%(91.9–95.7)	20.6% (16.8–25.0)	35.3% (30.6–40.3)	50.4% (45.3–55.5)	69.9% (65.0–74.3)	85.7%(81.7–88.9)
PVpos	1.2%(0.6–2.4)	1.6%(0.9–3.0)	2.0%(1.2–3.4)	1.3% (0.6–2.5)	2.7%(1.7–4.3)	31.4% (26.9–36.3)	33.6% (29.0–38.5)	37.4% (32.6–42.5)	43.1% (38.1–48.1)	51.3% (46.2–56.3)
PVneg	99.3%(98.2–99.7)	99.5%(98.5–99.8)	99.3%(98.3–99.7)	99.1%(97.9–99.6)	99.1%(98.0–99.6)	98.2%(96.3–99.2)	89.7% (86.2–92.4)	86.5% (82.6–89.6)	83.3% (79.2–86.8)	79.8% (75.4–83.6)
Efficiency	45.9%(42.0–49.9)	60.0%(56.1–63.8)	75.4%(71.8–78.7)	86.2% (83.3–88.7)	93.3%(91.0–95.0)	41.7% (36.8–46.7)	49.7% (44.7–54.8)	58.2% (53.1–63.1)	67.7% (62.8–72.3)	73.7% (69.0–77.9)
2 × 2 table	2 FN 276 TN	2 FN 362 TN	3 FN 457 TN	5 FN 525 TN	5 FN 568TN	1 FN 56 TN	11 FN 96 TN	21 FN 134 TN	38 FN 190 TN	59 FN 233 TN
	4 TP 328 FP	4 TP 242 FP	3 TP 147 FP	1 TP 79 FP	1 TP 36 FP	99 TP 216 FP	89 TP 176 FP	79 TP 132 FP	62 TP 82 FP	41 TP 39 FP
tPSA concentration intervals	> 4 – ≤ 20 μg/L tPSA N = 512					> 20 μg/L tPSA N = 139
Cut-off %fPSA	30%^a	25%^a	20%^a	15%^a	10%^a	30%^a	25%^a	20%^a	15%^a	10%^a
Sensitivity	98.3% (96.8–99.1)	89.3% (86.4–91.7)	79.2% (75.5–82.5)	61.8% (57.5–65.9)	39.0% (34.9–43.3)	84.3% (77.4–89.5)	80.0% (72.6–85.8)	73.9% (66.0–80.5)	57.4% (49.1–65.3)	39.1% (31.4–47.4)
Specificity	19.2% (16.0–22.8)	33.5% (29.6–37.7)	49.1% (44.8–53.4)	68.3% (64.1–72.1)	84.3% (80.9–87.2)	12.5% (8.0–19.0)	20.8% (14.9–28.3)	37.5% (29.9–45.8)	58.3% (50.0–66.2)	79.2% (71.7–85.1)
PVpos	39.3% (35.2–43.6)	41.7% (37.5–46.0)	45.3% (41.1–49.7)	50.9% (46.6–55.2)	56.6% (52.2–60.8)	82.2% (75.0–87.7)	82.9% (75.8–88.2)	85.0% (78.1–90.0)	86.8% (80.2–91.5)	90.0% (83.9–94.0)
PVneg	95.5% (93.4–97.0)	85.5% (82.2–88.3)	81.6% (78.0–84.7)	77.0% (73.2–80.5)	72.4% (68.4–76.1)	14.3% (9.4–21.1)	17.9% (12.4–25.1)	23.1% (16.9–30.7)	22.2% (16.1–29.8)	21.3% (15.4–29.9)
Efficiency	46.7% (42.4–51.0)	52.9% (48.6–57.2)	59.6% (55.3–63.7)	66.0% (61.8–70.0)	68.7% (64.5–72.5)	71.9% (64.0–78.7)	69.8% (61.7–76.8)	67.6% (59.5–74.8)	57.6% (49.2–65.5)	46.0% (38.0–54.3)
2 × 2 table	3 FN 64 TN	19 FN 112 TN	37 FN 164 TN	68 FN 228 TN	108 FN 284TN	18 FN 3 TN	23 FN 5 TN	30 FN 9 TN	49 FN 14 TN	70 FN 19 TN
	175 TP 270 FP	159 TP 222 FP	141 TP 170 FP	110 TP 106 FP	69 TP 53 FP	97 TP 21 FP	92 TP 19 FP	85 TP 15 FP	66 TP 10 FP	45 TP 5 FP

A total of 1261 patients with PSA measurements were eligible for the study; 299 patients were diagnosed with Prostate cancer (PCa), and 962 patients were found without PCa. %fPSA denotes percentage free Prostate Specific Antigen relative to tPSA. tPSA denotes total Prostate Specific Antigen. The cut-offs for %fPSA were set to 30, 25, 20, 15 and 10%. The tPSA concentration intervals were set to ≤4 μg/L, > 4 – ≤ 10 μg/L, > 4 – ≤ 20 μg/L, and >20 μg/L. FN: False Negative patients; TN: True Negative patients; TP: True Positive patients; FP: False Positive patients; PVpos: Predictive Value of a positive test; PVneg: Predictive Value of a negative test. Sensitivity was defined as TP/(TP + FN). Specificity was defined as TN/(TN + FP). PVpos was defined as TP/(TP + FP). PVneg was defined as TN/(TN + FN). Efficiency was defined as (TP + TN)/TP + FN + TN + FP.

^aConcentrations ≥ the specified cut-off were classified as FN (PCa) and TN (non-PCa) patients, respectively. Concentrations < the specified cut-off were classified as TP (PCa) and FP (non-PCa) patients, respectively. 95% confidence intervals are shown in italics.

Table 4. Multivariate Cox Regression Analysis for overall survival and PCa specific survival. Hazard ratios (range).

	All patients (N = 1261)		PCa patients (N = 299)
Variable	Overall survival	PCa specific survival	Overall survival	PCa specific survival
Age	2.247*** (2.005–2.518)	1.463* (1.079–1.983)	1.693*** (1.364–2.100)	1.348 (0.988–1.839)
tPSA	1.036 (0.921–1.165)	2.107*** (1.503–2.955)	1.439*** (1.161–1.782)	1.774** (1.239–2.541)
Cancer status	1.401* (1.078–1.820)	21.215*** (6.158–73.084)	NE	NE
Gleason Score	NE	NE	1.315** (1.096–1.579)	1.704*** (1.240–2.343)

PCa: Prostate cancer. Hazard ratio denotes the probability of death in one group compared with the probability of death in another group. Age categories: < 50, 50–60, 60–70, 70–80 or >80 years. tPSA categories: <4 μg/L, 4–10 μg/L, 10–20 μg/L, >20 μg/L. Cancer status: PCa or non-PCa. Gleason Score: 2–5, 6, 7 or 8–10. 95% Confidence intervals are shown in italics. NE: not eligible.

* p < .05,

** p < .005,

*** p < .001.

Figure 3. Kaplan-Meier plot of 10-year overall survival (A) and 10-year PCa-specific survival in different tPSA concentration intervals (B). All patients both prostate cancer and non-prostate cancer patients are included in each tPSA concentration interval. PCa denotes patients with prostate cancer. tPSA denotes total Prostate Specific Antigen.