Articles

Multiple Events of Fractures and Cardiovascular and Thromboembolic Disease Following Prostate Cancer Diagnosis: Results From the Population-Based PCBaSe Sweden

By: Mieke Van Hemelrijcka lowast , Hans Garmoa b, Lars Holmberga b, Pär Stattinc and Jan Adolfssond

Published online: 01 April 2012

Keywords: Prostate cancer, Thromboembolic disease, Endocrine treatment, Curative treatment, Surveillance, Cardiovascular disease, Fracture

Abstract Full Text Full Text PDF (270 KB)

Abstract

Background

To date, adverse events of prostate cancer (PCa) treatment have only been studied as a single event, and little is known about the risk of subsequent adverse events.

Objective

We assessed the frequency of multiple events (fractures, stroke, heart disease [HD], and thromboembolic disease [TED]) following PCa diagnosis.

Design, setting, and participants

PCBaSe Sweden is based on the National Prostate Cancer Register (NPCR) that covers >96% of incident PCa cases in Sweden.

Measurements

We evaluated the number of events (fractures, stroke, HD, and TED) leading to hospitalisation recorded in the National Hospital Discharge Registry after PCa diagnosis and conducted multivariate age-adjusted Cox proportional hazards regression to estimate the risk of developing multiple events.

Results and limitations

Between 1997 and 2007, 30 642 men received primary endocrine treatment, 26 432 curative treatment, and 19 526 surveillance: 75% had no event during follow-up, 17% had one event, and 9% had more than one event. The incidence of any event was 102 in 1000 person-years. Men who already had experienced an event, particularly HD, before or after the date of PCa diagnosis were more likely to have multiple events afterwards. For example, the hazard ratio of developing a third event for those with two or more events of HD before PCa diagnosis was 1.40 (95% confidence interval, 1.28–1.52) compared with those with no events of HD before PCa diagnosis. Events treated without hospitalisation were not included, so the number of adverse events is possibly underestimated.

Conclusions

A third of PCa patients with an adverse event after treatment subsequently experienced another adverse event, but apart from history of HD or stroke before PCa diagnosis, no specific characteristics were found for these men. Thus PCa management needs to take into account the risk of adverse events in all PCa patients, especially those with a history of adverse events before PCa diagnosis.

Take Home Message

Prostate cancer (PCa) management needs to take into account the risk of adverse events in all PCa patients, especially those on endocrine treatment or with a history of adverse events before PCa diagnosis.

Keywords: Prostate cancer, Thromboembolic disease, Endocrine treatment, Curative treatment, Surveillance, Cardiovascular disease, Fracture.

1. Introduction

Endocrine treatment (ET), which interrupts testosterone stimulation of the prostate tumour, is the gold standard for treatment of metastatic prostate cancer (PCa). Use of ET has increased over time and is currently also used in a substantial proportion of men with localised PCa, resulting in more men on hormonal treatment for longer times [1].

A number of metabolic side effects have been reported including increased body weight, hyperglycaemia, and osteopenia [2], [3], [4], and [5]. Based on several studies [3], [6], [7], [8], [9], [10], and [11], a safety announcement of the US Food and Drug Administration requires the manufacturers of gonadotropin-releasing hormone (GnRH) agonists to add new safety information about the risk of cardiovascular disease to the “Warning and Precautions” section of drug labels [12]. For instance, a study based on 37 443 men diagnosed with local or regional PCa in the Veterans Healthcare Administration Programme showed that treatment with GnRH agonists was associated with statistically significantly increased risks of incident diabetes, incident coronary heart disease, myocardial infarction, sudden cardiac death, and stroke [6].

In addition to an increased risk of different types of nonfatal and fatal heart disease (HD) and stroke for PCa patients on GnRH agonists, we found an increased risk for HD in all men with PCa in PCBaSe Sweden [9]. We also showed that these men with PCa are at increased risk of thromboembolic diseases (TEDs) [13]. When comparing different treatment groups (ET, curative treatment, and surveillance), men on ET had the highest risk for developing both HD and TED [9] and [13].

Besides their role in the regulation of heart functioning, androgens stimulate bone growth. As a result, a rapid loss of bone-mineral density occurs within the first 12 mo of ET [14], [15], and [16]. A survival analysis based on data in the Surveillance Epidemiology and End Results program showed that among men surviving at least 5 yr after diagnosis, 19% of men on androgen-deprivation therapy had had a fracture, compared with 13% for men with PCa not on ET (p<0.001) [16]. In PCBaSe Sweden, the standardised incidence ratio for risk of any fracture was also increased (unpubl. data, Thorstenson et al).

Thus PCa patients are at increased risk of several types of adverse events, especially following ET. However, to date these adverse events have only been studied as a single event, and little is known about the risk of subsequent adverse events. Adverse events are also observed following other types of PCa treatment. For instance, surgical cancer treatments have been shown to increase the risk of TED [17]. In a cohort of 523 consecutive patients undergoing radical prostatectomy with lymphadenectomy, a significant proportion of patients developed deep vein thrombosis (DVT) after day 8 [18]. Another study based on 773 consecutive patients undergoing laparoscopic radical prostatectomy showed that pelvic lymph node dissection during radical prostatectomy increased the risk of DVT [19].

We assessed the distribution of multiple adverse events after PCa diagnosis posing the question whether a specific group of men is at increased risk of developing multiple adverse events such as HD, stroke, TED, and fractures.

2. Methods

The current study is based on the same cohort used to show that men with PCa are at an increased risk of HD, stroke, TED, and fractures compared with the total Swedish male population [9] and [13]. Those studies assessed the effect of different types of ET (GnRH agonists, orchiectomy, antiandrogen [AA] monotherapy, and combinations of different ET) and curative treatment (radical prostatectomy and radiotherapy), so that the current study has a specific focus on the development of multiple adverse events (HD, stroke, TED, and fractures).

2.1. Study population and data collection

PCBaSe Sweden is based on the National Prostate Cancer Register (NPCR) of Sweden, which started in 1996 and captures >96% of all incident, biopsy-confirmed PCa as compared with the Swedish Cancer Registry, which has an underreporting for PCa <3.7% [20]. The NPCR includes date of diagnosis, age at diagnosis, tumour stage, tumour differentiation, serum level of prostate-specific antigen (PSA) at the time of diagnosis, and primary treatment given or planned up to 6 mo after the date of diagnosis. The validity of primary treatment registered in NPCR is >90% for curative treatment and surveillance, and >95% for ET (P. Stattin, unpublished data). A more detailed description of NPCR is given elsewhere [21]. By using the Swedish 10-digit personal identity number, PCBaSe could be linked to other national registers, allowing for information to be gathered regarding demographics, comorbidities, socioeconomic status (SES), and causes of death [22] and [23]. As of 1987, the Hospital Discharge Register collects information regarding inpatient care. Each record contains medical information on surgical and anaesthetic procedures, hospital department, and discharge diagnoses coded according to the International Classification of Diseases, 10th revision (ICD-10) [22]. For heart diseases, the primary diagnosis has been shown to be correct for about 95% of patients as judged by the European Society of Cardiology diagnosis guidelines [24], [25], and [26]. SES characteristics were assessed by record linkages to the 1960–1990 5-yr census databases. SES is based on occupational group and stratifies men into white-collar worker, blue-collar worker, not gainfully employed, and unknown[23]. As of 1997, the Cause of Death Register collects information on date and underlying cause of death coded according to the ICD-10[22]. Detailed information on the data content of PCBaSe is given elsewhere [27].

For our analyses the following information was taken from PCBaSe: age; serum levels of PSA and treatment information at time of diagnosis; tumour grade and stage; SES; history of TED, HD, stroke, or fractures; and date of death. Information on history of events involved primary diagnoses as registered in the Hospital Discharge Register. Gleason score was used to assess tumour grade. If World Health Organisation (WHO) tumour grade was reported primarily instead of Gleason score (25% of men), conversion to Gleason score was conducted as follows: G1=Gleason 2–6, G2=Gleason 7, and G3=Gleason 8–10. PCa stage was defined based on the TNM stages used in the NPCR [21]. Men with PCa were selected if they received curative treatment, surveillance, or ET as primary treatment. The curative treatment group consisted of men who underwent radical prostatectomy and/or radiotherapy; the surveillance group consisted of men who were on active surveillance or watchful waiting [28].

2.2. Analysis

We studied the following types of adverse event as primary diagnoses in the Hospital Discharge Register: HD (ICD-10 codes I20–I25), acute myocardial infarction (ICD-10 code I21), arrhythmia (ICD-10 codes I44–I49), heart failure (ICD-10 code I50), stroke (ICD-10 codes I60–64, G45), TED (DVT; ICD-10 codes: I80–82), pulmonary embolism (ICD-10 code I26), arterial embolism (ICD-10 codes K55, I74), and fractures (ICD-10 codes S12, S22, S32, S42, S52, S62, S72, S82, and S92). These different diseases are denoted later as “events.”

To compare the distribution of multiple adverse events between men with and without PCa, we also randomly selected a comparison cohort of three men free of PCa of the same age (±1 yr) from the same county for each PCa patient registered in PCBaSe between January 1, 1997, and December 31, 2006. Thus for each man with PCa, our study contains three matched men without PCa, denoted later as the “comparison cohort.”

We first showed the number of events by PCa treatment. To take into account the possible lag time between treatment and events, the number of events was also assessed while excluding those events occurring within 1, 3, and 6 mo after diagnosis. Descriptive statistics were also used to identify whether there were any differences in demographic or tumour characteristics by number of events after PCa diagnosis. We also modelled the risk of getting multiple events with a multivariable age-adjusted Cox proportional hazards model among those men with PCa who had already experienced at least one event after PCa diagnosis. More specifically, the risk of developing a second event was modelled among those who had already experienced one event after PCa diagnosis, whereas the risk of developing a third event was calculated among those who had already experienced a second event, and the risk of developing a fourth event was modelled among those who had already experienced a third event after PCa diagnosis. Due to small numbers, we did not model the risk of getting more than four events.

Incidence of different events was plotted for different groups of PCa patients according to their history of events. These incidences refer to number of events, whereby each person can contribute several events. For both men with PCa and their comparison cohort, we then showed the proportion of men still alive who experienced various combinations of events after PCa diagnosis. These proportions were calculated at fixed time points of follow-up (from 0 to 0.5, 1, 1.5, 2, and every 6 mo to 5 yr) among men still alive at these time points. To illustrate the intensity of events occurring, we also plotted the cumulative incidence of death among men with PCa and their comparison cohort.

Statistical analyses were performed with SAS v.9.1.3 (SAS Institute, Cary, NC, USA) and R v.2.7.2 (R Foundation for Statistical Computing, Vienna, Austria). The Swedish Central Ethics Committee (Dnr Ö 14-2007) and the Ethics Committee at Umeå University (Dnr 07-049M) approved the project.

3. Results

Between 1997 and 2007, NPCR registered 76 600 men diagnosed with PCa of whom 30 642 were treated with ET as their primary treatment. Specifically, 3391 men received AA; 5340 underwent orchiectomy; 9066 received GnRH agonists; and 11 646 men received GnRH agonists combined with short-time AA. The remaining 1199 men were treated with other types or combinations of ET. Another 45 958 men received curative treatment (26 432; 58%) or surveillance (19 525; 42%). Table A.1 in the appendix provides a more detailed description of the study population.

After a mean follow-up time of 4 yr after PCa diagnosis, 12 806 men (17%) had experienced one event (ie, HD, stroke, TED, or fracture), and 6528 men (9%) had had more than one event (Table 1). Most of the events occurred in men on ET (49%) and in men on surveillance (31%). TED occurred among 2% of PCa patients, whereas fractures occurred among 5%, stroke among 6%, and HD among 14% of the patients. For each type of adverse event, most patients only developed one event (Table 1). When excluding those events occurring within 1 mo after PCa diagnosis, 12 459 men (16%) had experienced one event and 6135 men (8%) had had more than one event. A similar pattern with fewer events was seen when excluding events occurring within 3 or 6 mo after PCa diagnosis (12 748 and 5002 or 11 127 and 3793, respectively). Most of the events lost when excluding the first months after PCa diagnosis are HD and stroke (results not shown).

Table 1 Frequency of events (heart disease, stroke, thromboembolic disease, and fractures) by prostate cancer treatment group

Endocrine treatment Surveillance Radical prostatectomy Radiotherapy All men with PCa Comparison cohort p value/Dispersion parameter
Any event (HD, stroke, TED, fractures), n (%) 0.0003/2.04
No events 21 125 (68.9) 13 591 (69.6) 12 466 (88.9) 7069 (82.0) 57 266 (74.8) 59 856 (78.1)
Only one event 6407 (20.9) 3658 (18.7) 1134 (8.1) 1069 (12.4) 12 806 (16.7) 11 001 (14.4)
Two events 1988 (6.5) 1386 (7.1) 263 (1.9) 292 (3.4) 4079 (5.3) 3485 (4.5)
Three events 667 (2.2) 506 (2.6) 88 (0.6) 110 (1.3) 1421 (1.9) 1308 (1.7)
More than three events 455 (1.5) 385 (2.0) 75 (0.5) 78 (0.9) 1028 (1.3) 950 (1.2)
HD, n (%) 0.51/2.29
No HD 25 685 (83.8) 16201 (83.0) 13 867 (98.9) 8468 (98.3) 66 011 (86.2) 66 912 (87.4)
Only one HD 2936 (9.6) 1788 (9.2) 117 (0.8) 105 (1.2) 6192 (8.1) 5620 (7.3)
More than one HD and no other events 1325 (4.3) 1078 (5.5) 11 (0.1) 14 (0.2) 3039 (4.0) 2864 (3.7)
One HD and one other event 568 (1.9) 362 (1.9) 25 (0.2) 15 (0.2) 1098 (1.4) 936 (1.2)
One HD and two other events 105 (0.3) 69 (0.4) 2 (0.0) 12 (0.1) 205 (0.3) 218 (0.3)
One HD and more than two other events 23 (0.1) 28 (0.1) 4 (0.0) 4 (0.0) 55 (0.1) 50 (0.1)
Stroke, n (%) 0.14/1.67
No stroke 28 361 (92.6) 17845 (91.4) 13 673 (97.5) 8232 (95.5) 71 721 (93.6) 72 093 (94.1)
Only one stroke 1432 (4.7) 980 (5.0) 261 (1.9) 263 (3.1) 3053 (4.0) 2849 (3.7)
More than one stroke and no other events 194 (0.6) 174 (0.9) 26 (0.2) 30 (0.3) 436 (0.6) 453 (0.6)
One stroke and one other event 441 (1.4) 327 (1.7) 41 (0.3) 51 (0.6) 896 (1.2) 784 (1.0)
One stroke and two other events 131 (0.4) 124 (0.6) 17 (0.1) 23 (0.3) 299 (0.4) 251 (0.3)
One stroke and more than two other events 83 (0.3) 76 (0.4) 8 (0.1) 19 (0.2) 195 (0.3) 170 (0.2)
TED, n (%) <0.0001/1.52
No TED 29 851 (97.4) 19 115 (97.9) 13 069 (93.2) 7704 (89.4) 75 011 (97.9) 75 736 (98.9)
Only one TED 497 (1.6) 239 (1.2) 626 (4.5) 571 (6.6) 1011 (1.3) 529 (0.7)
More than one TED and no other events 38 (0.1) 18 (0.1) 274 (2.0) 260 (3.0) 83 (0.1) 49 (0.1)
One TED and one other event 160 (0.5) 84 (0.4) 48 (0.3) 68 (0.8) 300 (0.4) 159 (0.2)
One TED and two other events 57 (0.2) 38 (0.2) 8 (0.1) 13 (0.2) 114 (0.1) 78 (0.1)
One TED and more than two other events 39 (0.1) 32 (0.2) 1 (0.0) 2 (0.0) 81 (0.1) 49 (0.1)
Fractures, n (%) <0.000
No fractures 28 280 (92.3) 18 450 (94.5) 13 860 (98.8) 8419 (97.7) 72641 (94.8) 73 596 (96.1)
Only one fractures 1542 (5.0) 651 (3.3) 130 (0.9) 130 (1.5) 2550 (3.3) 2003 (2.6)
More than one fracture and no other events 185 (0.6) 90 (0.5) 4 (0.0) 13 (0.2) 303 (0.4) 202 (0.3)
One fractures and one other event 431 (1.4) 201 (1.0) 20 (0.1) 38 (0.4) 722 (0.9) 511 (0.7)
One fractures and two other events 120 (0.4) 78 (0.4) 8 (0.1) 12 (0.1) 228 (0.3) 164 (0.2)
One fractures and more than two other events 84 (0.3) 56 (0.3) 4 (0.0) 6 (0.1) 156 (0.2) 124 (0.2)

PCa=prostate cancer; HD=heart disease; TED=thromboembolic disease.

We then assessed whether there are any specific patient or tumour characteristics for those 9% of men who developed more than one adverse event. Table 2 shows that multiple adverse events occurred after all types of primary events, but more men had an event of HD before PCa diagnosis in the group with two or three or more events than those with one event (eg, 36%, 48%, and 28%, respectively, had two or more events of HD before PCa diagnosis). This observation is also seen in Table 3, where we show the results of age-adjusted multivariable models for the risk of developing a second, third or fourth event. Those who had experienced two or more events of HD before PCa diagnosis were 1.4 times (95% confidence interval [CI], 1.28–1.52) more likely to develop a third event compared with those who were free of HD before PCa diagnosis. These multivariable models also show that history of stroke is related to an increased risk of developing a third event (hazard ratio [HR]: 1.51; 95% CI, 1.20–1.89). Those who develop HD after PCa diagnosis are also more likely to develop multiple events of HD after PCa diagnosis compared with those who develop a fracture after PCa diagnosis (eg, the HR of developing a fourth event after having had three events of HD is 3.55 (95% CI, 1.58–7.99). Table 3 shows that when radical prostatectomy is considered the reference group, there is no strong difference in risk of developing multiple events by treatment. All other treatment are associated with an increased risk of a second event, apart from radiotherapy and other curative treatment (eg, HR of developing a second event after having had a first event is 1.32 [95% CI, 1.18–1.48] for those on ET and 1.27 [95% CI, 1.13–1.41] for those on surveillance) (Table 3).

Table 2 Demographic and tumour characteristics by number of events (heart disease, stroke, thromboembolic disease, or fractures) after prostate cancer diagnosis

No event after PCa diagnosis One event after PCa diagnosis Two events after PCa diagnosis Three or more events after PCa diagnosis
Age, yr, n (%)
<65 16 924 (29.6) 1575 (12.3) 381 (9.3) 242 (9.9)
65–74 21 245 (37.1) 4306 (33.6) 1357 (33.3) 841 (34.3)
≥75 19 097 (33.3) 6925 (54.1) 2341 (57.4) 1366 (55.8)
Stage, n (%)
Localised: PSA <20 28 433 (49.7) 4926 (38.5) 1659 (40.7) 1009 (41.2)
Localised: 20 ≤ PSA <50 5569 (9.7) 1527 (11.9) 518 (12.7) 361 (14.7)
Locally advanced 6720 (11.7) 1897 (14.8) 613 (15.0) 389 (15.9)
Intermediate group 4448 (7.8) 1339 (10.5) 404 (9.9) 257 (10.5)
Metastatic disease 9695 (16.9) 2512 (19.6) 671 (16.5) 309 (12.6)
Missing data 2401 (4.2) 605 (4.7) 214 (5.2) 124 (5.1)
Primary treatment, n (%)
Endocrine treatment 20 971 (36.6) 6356 (49.6) 1974 (48.4) 1119 (45.7)
Other/unknown palliative 997 (1.7) 301 (2.4) 92 (2.3) 53 (2.2)
Radical prostatectomy 12 466 (21.8) 1134 (8.9) 263 (6.4) 163 (6.7)
Radiotherapy 7069 (12.3) 1069 (8.3) 292 (7.2) 188 (7.7)
Other/unknown curative treatment 422 (0.7) 61 (0.5) 7 (0.2) 8 (0.3)
Surveillance 13 591 (23.7) 3658 (28.6) 1386 (34.0) 891 (36.4)
Missing data/other 1750 (3.1) 227 (1.8) 65 (1.6) 27 (1.1)
No. of patients with HD before PCa diagnosis (%)
0 48 284 (84.3) 9287 (72.5) 2613 (64.1) 1274 (52.0)
1 3987 (7.0) 1416 (11.1) 539 (13.2) 358 (14.6)
≥2 4994 (8.7) 2103 (16.4) 927 (22.7) 817 (33.4)
No. of strokes before PCa diagnosis (%)
0 53 635 (93.7) 11 381 (88.9) 3574 (87.6) 2120 (86.6)
1 2486 (4.3) 963 (7.5) 326 (8.0) 213 (8.7)
≥2 1145 (2.0) 462 (3.6) 179 (4.4) 116 (4.7)
No. of patients with TED before PCa diagnosis (%)
0 56 086 (97.9) 12 357 (96.5) 3917 (96.0) 2334 (95.3)
1 929 (1.6) 346 (2.7) 128 (3.1) 88 (3.6)
≥2 251 (0.4) 103 (0.8) 34 (0.8) 27 (1.1)
No. of fractures before PCa diagnosis (%)
0 54 280 (94.8) 11933 (93.2) 3775 (92.5) 2294 (93.7)
1 2619 (4.6) 752 (5.9) 259 (6.3) 135 (5.5)
≥2 367 (0.6) 121 (0.9) 45 (1.1) 20 (0.8)

PCa=prostate cancer; PSA=prostate-specific antigen; HD=heart disease; TED=thromboembolic disease.

Table 3 Hazard ratios and 95% confidence intervals for developing a second, third, and fourth event*

No. of events/No. of men with PCa at risk of developing event Second event Third event Fourth event
6967/19 372 2770/6916 1242/2756
HR 95% CI HR 95% CI HR 95% CI
Treatment
Radical prostatectomy 1.00 Ref. 1.00 Ref. 1.00 Ref.
Endocrine treatment 1.32 (1.18–1.48) 1.07 (0.90–1.28) 0.92 (0.71–1.19)
Other palliative treatment 1.21 (1.00–1.46) 1.03 (0.76–1.38) 1.19 (0.77–1.86)
Radiotherapy 1.04 (0.92–1.19) 0.96 (0.78–1.17) 0.90 (0.67–1.22)
Other curative treatment 0.71 (0.43–1.17) 1.72 (0.88–3.37) 1.32 (0.48–3.60)
Surveillance 1.27 (1.13–1.41) 0.98 (0.82–1.16) 0.95 (0.73–1.23)
Missing 1.38 (1.10–1.72) 0.94 (0.64–1.39) 1.22 (0.71–2.07)
Type of events after PCa diagnosis
Only FRAC 1.00 Ref. 1.00 Ref. 1.00 Ref.
Only HD 1.78 (1.65–1.91) 2.53 (1.98–3.25) 3.55 (1.58–7.99)
Only stroke 1.10 (1.00–1.20) 1.50 (1.13–2.01) 2.00 (0.83–4.80)
Only TED 1.15 (1.01–1.30) 1.81 (1.19–2.74) 1.57 (0.32–7.85)
Mixture 1.83 (1.42–2.35) 2.54 (1.13–5.73)
FRAC before PCa diagnosis
0 1.00 Ref. 1.00 Ref. 1.00 Ref.
1 1.09 (0.99–1.21) 0.94 (0.80–1.10) 1.19 (0.94–1.51)
≥2 1.26 (0.99–1.60) 0.92 (0.62–1.37) 1.53 (0.84–2.79)
HD before PCa diagnosis
0 1.00 Ref. 1.00 Ref. 1.00 Ref.
1 1.25 (1.17–1.34) 1.28 (1.15–1.43) 1.04 (0.88–1.23)
≥2 1.52 (1.44–1.61) 1.40 (1.28–1.52) 1.30 (1.15–1.48)
Stroke before PCa diagnosis
0 1.00 Ref. 1.00 Ref. 1.00 Ref.
1 1.16 (1.07–1.25) 1.18 (1.04–1.33) 1.23 (1.03–1.46)
≥2 1.28 (1.10–1.49) 1.51 (1.20–1.89) 1.63 (1.18–2.27)
TED before PCa diagnosis
0 1.00 Ref. 1.00 Ref. 1.00 Ref.
1 1.19 (1.04–1.36) 1.24 (1.02–1.51) 0.99 (0.74–1.33)
≥2 1.10 (0.86–1.42) 1.30 (0.90–1.87) 1.65 (1.03–2.63)

* All HRs are taken from an age-adjusted multivariable Cox proportional hazards regression model.

PCa=prostate cancer; HR=Hazard ratio; CI=confidence interval; FRAC=fracture; HD=heart disease; TED=thromboembolic disease.

To take into account the time since date of PCa diagnosis, we plotted the number of events per person-time by follow-up time: The incidence of any adverse event was 102 per 1000 person-years. Figure 1 shows the incidences of an adverse event after HD, stroke, TED, or a fracture after PCa diagnosis. The incidence of HD was always highest, and the second most common adverse event was always the same kind of event that had already occurred after PCa diagnosis, with the exception of TED.

gr1

Fig. 1 Incidence (per 1000 person-years) of events of heart disease (HD), stroke, thromboembolic disease (TED), or fracture after prostate cancer (PCa) diagnosis for PCa patients with one previous event of HD, stroke, fracture or TED, respectively, after PCa diagnosis.

Finally, we compared the occurrence of multiple events between men with PCa and their comparison cohort. Figure 2 shows the proportion of men alive who experienced different combinations of events after time of PCa diagnosis for men with PCa (Fig. 2a) and their comparison cohort (Fig. 2b). From this figure, three types of information can be derived: (1) The slope of the curve indicates how fast the number of events occurred, (2) the total number of events observed is on the y-axis, and (3) the proportion of different events is shown by different colours. Small differences were seen between men with PCa and their comparison cohort for the proportions of different event types (HD, stroke, TED, or fractures). However, when observing these proportions, it is important to take into account the fact that men with PCa die at a faster rate than their comparison cohort (Fig. 2c), so the intensity at which these multiple events happen is much larger for those with PCa than those in the comparison cohort.

gr2

Fig. 2 (a) Proportions of men still alive who experienced various combinations of events (heart disease [HD], stroke, thromboembolic disease [TED], or fracture) after being diagnosed with PCa and (b) their comparison cohort. Three types of information can be observed: (1) the slope of the curve indicates how fast the number of events is increasing, (2) the total number of events is observed on the y-axis, and (3) the proportion of different event types is shown by different colours. (c) The cumulative incidence of death among men with PCa and their comparison cohort then indicates that the intensity of events (HD, stroke, TED, or fracture) must be larger for those with PCa.

4. Discussion

This is the first study describing the risk of subsequent cardiovascular events or fractures following PCa diagnosis. About 25% of PCa patients experienced an event including HD, stroke, or fracture, and about 33% of these men subsequently had one or more events. Apart from a history of HD before PCa diagnosis, no particular characteristics were found for men who experienced multiple events. The incidence of any event after PCa diagnosis was 102 in 1000 person-years.

Information on adverse events was taken from the National Hospital Discharge Register. Thus milder adverse events not requiring hospitalisation were not included, which likely resulted in an underestimation of the number of adverse events. It is also possible that some adverse events are underestimated due to “early death.” However, because this study has a clinical rather than explanatory focus, we believe that competing risks are not a major issue. In addition, the mean follow-up time for this study was 4 yr, which is probably not long enough to capture all adverse events following PCa treatment. The diagnosis of PCa itself can bias the results because these men receive more intensive medical attention. However, the similar pattern for men with PCa and their control group (Fig. 2) indicates there was no detection bias following close surveillance in PCa patients. We had no information about smoking habits, diabetes, body mass index, or hypertension, which may be associated with choice of treatment for PCa patients. Even though we ran age-adjusted multivariable models with detailed information on stage of disease, treatment, and history of adverse events, there is residual bias that cannot be accounted for. Choice of ET, for instance, is likely to be related to comorbidity, suggesting there could be a selection bias remaining. However, the physician's and/or patient's choice of ET primarily constitutes a confounder for studying the intended effect (palliative treatment for PCa) but not for studying adverse effects that were not known at the time of data collection [29]. It was not standard clinical practice to take HD history into account when initiating ET in Sweden at the time. In our previous studies we already showed that a possible delay between PCa diagnosis and start of ET was unlikely to affect our results [9] and [13]. Treatment in this study was based on primary treatment given or planned up to 6 mo after the date of diagnosis, and no information on treatment dose or continuation was available. Finally, an unknown proportion of men curatively treated or on surveillance changed to ET, which could dilute a true difference in risk between different treatment groups. Therefore, this study could not assess in detail the effect of treatment duration or intermittency.

Several studies to date have shown that PCa patients are at increased risk of several types of adverse events, especially following ET [2], [3], [4], [5], [7], [8], and [30]. Recently, there has been an increasing interest in risk of HD following ET on the basis of a theoretically assumed reduced cardioprotective effect of testosterone [3], [7], [8], [30], and [31]. For example, a study based on 1372 men showed that the use of ET was associated with earlier onset of fatal myocardial infarction in men ≥65 yr of age who were treated for 6 mo compared with men who were not treated with ET [30]. However, all these studies assessed the risk of these adverse events by focusing on a single event, and little is known about the risk of subsequent adverse events. In addition to ET, several studies have investigated the risk of TED following PCa surgery. Radical prostatectomy is known to be associated with the risk of venous thromboembolism, and it is included in the list of urologic procedures for which pharmacologic TED prophylaxis is recommended, even for patients discharged early postoperatively [32]. For instance, in a cohort study of 523 consecutive patients undergoing radical prostatectomy with lymphadenectomy, 17% developed TED [18].

The largest number of events occurred among men treated with ET (Table A.1) followed by men on surveillance, and men treated with curative treatment had the lowest incidence of adverse events. This pattern indicates that treatments and selection to treatment both affect the risk for an adverse event. Men with comorbidities are more likely to be put on surveillance or treatment with noncurative intent, whereas curative treatments are more likely to be given to healthier men [33]. Stroke and HD, for example, occurred mostly among men on surveillance, whereas TED and fractures occurred more often among men on ET. Most multiple events occurred among men who first developed one HD or stroke, as expected, because these are chronic, recurrent, and progressive diseases [28] and [29].

A prior history of HD or stroke identifies men at moderately increased risk of multiple adverse events. A high prevalence of HD incidence also indicates an elevated risk for HD, stroke, TED, or fracture in men who already had experienced an event after PCa diagnosis. We could not substantiate that any specific demographic or tumour characteristic signifies those who develop multiple events of HD, stroke, TED, or fractures after PCa diagnosis. Thus these data contradict the hypothesis that only a smaller group of vulnerable men experience all types of adverse effects reported in different studies.

5. Conclusions

Our findings show that clinical information routinely available for PCa patients cannot identify men at increased risk of multiple adverse events with enough precision to be of help in designing surveillance programs. However, primary events of HD or stroke after PCa diagnosis were associated with an increased risk of multiple events of HD or stroke.

Author contributions: Mieke Van Hemelrijck had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Adolfsson, Van Hemelrijck, Holmberg, Garmo.

Acquisition of data: Stattin, Garmo.

Analysis and interpretation of data: Van Hemelrijck, Garmo.

Drafting of the manuscript: Van Hemelrijck.

Critical revision of the manuscript for important intellectual content: Holmberg, Garmo, Adolfsson, Stattin.

Statistical analysis: Garmo.

Obtaining funding: Holmberg, Stattin, Adolfsson.

Administrative, technical, or material support: Stattin.

Supervision: Holmberg, Adolfsson.

Other (specify): None.

Financial disclosures: I certify that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/ affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: None.

Funding/Support and role of the sponsor: Funding came from the Swedish Research Council 825-2008-5910, Stockholm Cancer Society, and Cancer Research-UK. This project was made possible by the continuous work of the National Prostate Cancer Register of Sweden (NPCR Swe) steering group: Pär Stattin chairman, Anders Widmark, Lars Egevad, Magnus Törnblom, Stefan Carlsson, Jan Adolfsson, Anna Bill-Axelson, Jan-Erik Johanssson, Ove Andreen, Mats Lambe, Erik Holmberg, David Robinson, Bill Pettersson, Jonas Hugosson, Jan-Erik Damber, Maria Nygren, Ola Bratt, and Göran Ahlgren.

See Table A1 and Table A2.

Table A.1 Baseline characteristics of men with prostate cancer according to their treatment in PCBaSe Sweden

Endocrine treatment Curative treatment Surveillance
Total ET group AAs Orchiectomy GnRH agonists GnRH agonists plus short AA
Total (%) 30 642 (100.0) 3391 (100.0) 5340 (100.0) 9066 (100.0) 11 646 (100.0) 26 432 (100.0) 19 526 (100.0)
Follow-up time (mean; SD) 3.5 (2.4) 4.0 (2.4) 3.1 (2.3) 3.8 (2.4) 3.3 (2.3) 4.4 (2.5) 4.7 (2.7)
Age group, yr (%)
<65 2941 (9.6) 546 (16.1) 208 (3.9) 731 (8.1) 1140 (9.8) 13 677 (51.7) 2535 (13.0)
65–74 9255 (30.2) 1505 (44.4) 1128 (21.1) 2706 (29.8) 3499 (30.0) 10 956 (41.4) 7526 (38.5)
75+ 18 446 (60.2) 1340 (39.5) 4004 (75.0) 5629 (62.1) 7007 (60.2) 1799 (6.8) 9465 (48.5)
Time period (%)
1997–1999 8503 (27.7) 534 (15.7) 2114 (39.6) 3113 (34.3) 2316 (19.9) 3849 (14.6) 5470 (28.0)
2000–2002 9669 (31.6) 1099 (32.4) 1507 (28.2) 2848 (31.4) 3850 (33.1) 6732 (25.5) 5618 (28.8)
2003–2006 12 470 (40.7) 1758 (51.8) 1719 (32.2) 3105 (34.2) 5480 (47.1) 15 851 (60.0) 8438 (43.2)
Gleason score (%)
2–6 6042 (19.7) 1016 (30.0) 851 (15.9) 2005 (22.1) 1961 (16.8) 15 325 (58.0) 13 277 (68.0)
7 12 192 (39.8) 1438 (42.4) 2069 (38.7) 3588 (39.6) 4690 (40.3) 8153 (30.8) 4687 (24.0)
8–10 11 562 (37.7) 877 (25.9) 2236 (41.9) 3205 (35.4) 4696 (40.3) 2729 (10.3) 1152 (5.9)
Missing data 846 (2.8) 60 (1.8) 184 (3.4) 268 (3.0) 299 (2.6) 225 (0.9) 410 (2.1)
Prostate Cancer Stage Group (%)
Localised: PSA<20 ng/ml 4298 (14.0) 838 (24.7) 469 (8.8) 1516 (16.7) 1348 (11.6) 18 850 (71.3) 12 879 (66.0)
Localised: 20 ≤ PSA <50 ng/ml 3317 (10.8) 516 (15.2) 460 (8.6) 1101 (12.1) 1158 (9.9) 2247 (8.5) 2411 (12.3)
Locally advanced 5323 (17.4) 643 (19.0) 829 (15.5) 1779 (19.6) 1931 (16.6) 2600 (9.8) 1696 (8.7)
Intermediate 4845 (15.8) 635 (18.7) 706 (13.2) 1561 (17.2) 1740 (14.9) 853 (3.2) 750 (3.8)
Metastatic disease or PSA >100 ng/ml 12 139 (39.6) 671 (19.8) 2726 (51.0) 2845 (31.4) 5269 (45.2) 592 (2.2) 456 (2.3)
Missing data 720 (2.3) 88 (2.6) 150 (2.8) 264 (2.9) 200 (1.7) 1290 (4.9) 1334 (6.8)
Civil status (%)
Married 19 630 (64.1) 2356 (69.5) 3202 (60.0) 5768 (63.6) 7488 (64.3) 19 451 (73.6) 13 379 (68.5)
Single 10 980 (35.8) 1033 (30.5) 2128 (39.9) 3291 (36.3) 4147 (35.6) 6972 (26.4) 6118 (31.3)
Missing data 32 (0.1) 2 (0.1) 10 (0.2) 7 (0.1) 11 (0.1) 9 (0.0) 29 (0.1)
Socioeconomic status
White collar 13 049 (42.6) 1749 (51.6) 1929 (36.1) 3839 (42.3) 4959 (42.6) 14 322 (54.2) 9203 (47.1)
Blue collar 17 105 (55.8) 1592 (46.9) 3334 (62.4) 5066 (55.9) 6504 (55.8) 11 805 (44.7) 10 052 (51.5)
Not gainfully employed/Missing data 488 (1.6) 50 (1.5) 77 (1.4) 161 (1.8) 183 (1.6) 305 (1.2) 271 (1.4)
TED at baseline (%)
Yes 1143 (3.7) 92 (2.7) 194 (3.6) 376 (4.1) 440 (3.8) 487 (1.8) 649 (3.3)
No 29 499 (96.3) 3299 (97.3) 5146 (96.4) 8690 (95.9) 11 206 (96.2) 25 945 (98.2) 18 877 (96.7)

ET=endocrine treatment; GnRH=gonadotropin-releasing hormone; AA=antiandrogen; SD=standard deviation; PSA=prostate-specific androgen; TED=thromboembolic disease.

Table A.2 Frequency of events (heart disease, stroke, thromboembolic disease, and fractures) by prostate cancer treatment group

Surveillance AA GnRH agonists GNRH agonists plus short-term AA Orchiectomy Other palliative treatment Radical prostatectomy Radiotherapy Other curative treatment Missing data All men with PCa
Any event (HD, stroke, TED, fractures)
No events 13591 (69.6) 2566 (75.7) 6588 (66.3) 8131 (69.8) 3686 (67.6) 997 (69.1) 12466 (88.9) 7069 (82.0) 422 (84.7) 1750 (84.6) 57266 (74.8)
Only one event 3658 (18.7) 580 (17.1) 2181 (22.0) 2430 (20.9) 1165 (21.4) 301 (20.9) 1134 (8.1) 1069 (12.4) 61 (12.2) 227 (11.0) 12806 (16.7)
Two events 1386 (7.1) 165 (4.9) 722 (7.3) 703 (6.0) 384 (7.0) 92 (6.4) 263 (1.9) 292 (3.4) 7 (1.4) 65 (3.1) 4079 (5.3)
Three events 506 (2.6) 37 (1.1) 266 (2.7) 227 (1.9) 135 (2.5) 31 (2.1) 88 (0.6) 110 (1.3) 6 (1.2) 15 (0.7) 1421 (1.9)
More than three events 385 (2.0) 43 (1.3) 177 (1.8) 155 (1.3) 79 (1.4) 22 (1.5) 75 (0.5) 78 (0.9) 2 (0.4) 12 (0.6) 1028 (1.3)
HD (%)
No HD 16201 (83.0) 2921 (86.1) 8170 (82.2) 9842 (84.5) 4567 (83.8) 1203 (83.4) 13069 (93.2) 7704 (89.4) 450 (90.4) 1884 (91.1) 66011 (86.2)
Only one HD 1788 (9.2) 294 (8.7) 1011 (10.2) 1102 (9.5) 502 (9.2) 141 (9.8) 626 (4.5) 571 (6.6) 36 (7.2) 121 (5.8) 6192 (8.1)
More than one HD and no other events 1078 (5.5) 118 (3.5) 499 (5.0) 480 (4.1) 222 (4.1) 62 (4.3) 274 (2.0) 260 (3.0) 6 (1.2) 40 (1.9) 3039 (4.0)
One HD and one other event 362 (1.9) 52 (1.5) 208 (2.1) 180 (1.5) 124 (2.3) 29 (2.0) 48 (0.3) 68 (0.8) 5 (1.0) 22 (1.1) 1098 (1.4)
One HD and two other events 69 (0.4) 6 (0.2) 39 (0.4) 32 (0.3) 28 (0.5) 7 (0.5) 8 (0.1) 13 (0.2) 1 (0.2) 2 (0.1) 205 (0.3)
One HD and more than two other events 28 (0.1) 0 (0.0) 7 (0.1) 10 (0.1) 6 (0.1) 1 (0.1) 1 (0.0) 2 (0.0) 0 (0.0) 0 (0.0) 55 (0.1)
Stroke (%)
No stroke 17845 (91.4) 3170 (93.5) 9127 (91.9) 10787 (92.6) 5072 (93.1) 1327 (92.0) 13673 (97.5) 8232 (95.5) 488 (98.0) 2000 (96.7) 71721 (93.6)
Only one stroke 980 (5.0) 150 (4.4) 487 (4.9) 562 (4.8) 222 (4.1) 73 (5.1) 261 (1.9) 263 (3.1) 8 (1.6) 47 (2.3) 3053 (4.0)
More than one stroke and no other events 174 (0.9) 17 (0.5) 69 (0.7) 72 (0.6) 33 (0.6) 9 (0.6) 26 (0.2) 30 (0.3) 0 (0.0) 6 (0.3) 436 (0.6)
One stroke and one other event 327 (1.7) 38 (1.1) 174 (1.8) 150 (1.3) 77 (1.4) 22 (1.5) 41 (0.3) 51 (0.6) 1 (0.2) 15 (0.7) 896 (1.2)
One stroke and two other events 124 (0.6) 7 (0.2) 51 (0.5) 41 (0.4) 31 (0.6) 5 (0.3) 17 (0.1) 23 (0.3) 0 (0.0) 0 (0.0) 299 (0.4)
One stroke and more than two other events 76 (0.4) 9 (0.3) 26 (0.3) 34 (0.3) 14 (0.3) 7 (0.5) 8 (0.1) 19 (0.2) 1 (0.2) 1 (0.0) 195 (0.3)
TED (%)
No TED 19115 (97.9) 3336 (98.4) 9650 (97.1) 11363 (97.6) 5286 (97.0) 1402 (97.2) 13867 (98.9) 8468 (98.3) 486 (97.6) 2038 (98.5) 75011 (97.9)
Only one TED 239 (1.2) 33 (1.0) 169 (1.7) 189 (1.6) 102 (1.9) 26 (1.8) 117 (0.8) 105 (1.2) 9 (1.8) 22 (1.1) 1011 (1.3)
More than one TED and no other events 18 (0.1) 1 (0.0) 18 (0.2) 9 (0.1) 9 (0.2) 2 (0.1) 11 (0.1) 14 (0.2) 0 (0.0) 1 (0.0) 83 (0.1)
One TED and one other event 84 (0.4) 11 (0.3) 55 (0.6) 61 (0.5) 32 (0.6) 9 (0.6) 25 (0.2) 15 (0.2) 3 (0.6) 5 (0.2) 300 (0.4)
One TED and two other events 38 (0.2) 7 (0.2) 28 (0.3) 11 (0.1) 11 (0.2) 3 (0.2) 2 (0.0) 12 (0.1) 0 (0.0) 2 (0.1) 114 (0.1)
One TED and more than two other events 32 (0.2) 3 (0.1) 14 (0.1) 13 (0.1) 9 (0.2) 1 (0.1) 4 (0.0) 4 (0.0) 0 (0.0) 1 (0.0) 81 (0.1)
Fractures (%)
No fractures 18450 (94.5) 3243 (95.6) 9113 (91.7) 10789 (92.6) 4926 (90.4) 1350 (93.6) 13860 (98.8) 8419 (97.7) 487 (97.8) 2004 (96.9) 72641 (94.8)
Only one fracture 651 (3.3) 103 (3.0) 514 (5.2) 577 (5.0) 339 (6.2) 61 (4.2) 130 (0.9) 130 (1.5) 8 (1.6) 37 (1.8) 2550 (3.3)
More than one fracture and no other events 90 (0.5) 7 (0.2) 69 (0.7) 66 (0.6) 42 (0.8) 7 (0.5) 4 (0.0) 13 (0.2) 1 (0.2) 4 (0.2) 303 (0.4)
One fracture and one other event 201 (1.0) 23 (0.7) 157 (1.6) 147 (1.3) 101 (1.9) 16 (1.1) 20 (0.1) 38 (0.4) 1 (0.2) 18 (0.9) 722 (0.9)
One fracture and two other events 78 (0.4) 7 (0.2) 47 (0.5) 44 (0.4) 22 (0.4) 7 (0.5) 8 (0.1) 12 (0.1) 1 (0.2) 2 (0.1) 228 (0.3)
One fracture and more than two other events 56 (0.3) 8 (0.2) 34 (0.3) 23 (0.2) 19 (0.3) 2 (0.1) 4 (0.0) 6 (0.1) 0 (0.0) 4 (0.2) 156 (0.2)

AA=antiandrogen; GnRH=gonadotropin-releasing hormone; PCa=prostate cancer; HD=heart disease; TED=thromboembolic disease.

References

  • [1] N. Sharifi, J.L. Gulley, W.L. Dahut. Androgen deprivation therapy for prostate cancer. JAMA. 2005;294:238-244 Crossref.
  • [2] M. Braga-Basaria, A.S. Dobs, D.C. Muller, et al. Metabolic syndrome in men with prostate cancer undergoing long-term androgen-deprivation therapy. J Clin Oncol. 2006;24:3979-3983
  • [3] N.L. Keating, A.J. O’Malley, M.R. Smith. Diabetes and cardiovascular disease during androgen deprivation therapy for prostate cancer. J Clin Oncol. 2006;24:4448-4456 Crossref.
  • [4] D. Mcleod, P. Iversen, W. See, T. Morris, J. Armstrong, M. Wirth. Bicalutamide 150mg plus standard care vs standard care alone for early prostate cancer. BJU Int. 2006;97:247-254 Crossref.
  • [5] Petrylak PJ, WM. Androgen ablation for prostate cancer: mechanisms and modalities. In: Kantoff PW, Carroll P, D’Amico D, editors. Prostate cancer: principles and practice. Philadelphia, PA: Lippincott Williams & Wilkins; 2002.
  • [6] N.L. Keating, A.J. O’Malley, S.J. Freedland, M.R. Smith. Diabetes and cardiovascular disease during androgen deprivation therapy: observational study of veterans with prostate cancer. J Natl Cancer Inst. 2010;102:39-46 Crossref.
  • [7] H.K. Tsai, A.V. D’Amico, N. Sadetsky, M.H. Chen, P.R. Carroll. Androgen deprivation therapy for localized prostate cancer and the risk of cardiovascular mortality. J Natl Cancer Inst. 2007;99:1516-1524 Crossref.
  • [8] C.S. Saigal, J.L. Gore, T.L. Krupski, J. Hanley, M. Schonlau, M.S. Litwin. Androgen deprivation therapy increases cardiovascular morbidity in men with prostate cancer. Cancer. 2007;110:1493-1500 Crossref.
  • [9] M. Van Hemelrijck, H. Garmo, L. Holmberg, et al. Absolute and relative risk of cardiovascular disease in men with prostate cancer: results from the population-based PCBaSe Sweden. J Clin Oncol. 2010;28:3448-3456 Crossref.
  • [10] S.M. Alibhai, M. Duong-Hua, R. Sutradhar, et al. Impact of androgen deprivation therapy on cardiovascular disease and diabetes. J Clin Oncol. 2009;27:3452-3458 Crossref.
  • [11] J.A. Efstathiou, K. Bae, W.U. Shipley, et al. Cardiovascular mortality after androgen deprivation therapy for locally advanced prostate cancer: RTOG 85-31. J Clin Oncol. 2009;27:92-99
  • [12] US Food and Drug Administration. FDA drug safety communication: update to ongoing safety review of GnRH agonists and notification to manufacturers of GnRH agonists to add new safety information to labeling regarding increased risk of diabetes and certain cardiovascular diseases. Rockville, MD: US Food and Drug Administration; 2010.
  • [13] M. Van Hemelrijck, J. Adolfsson, H. Garmo, et al. Risk of thromboembolic diseases in men with prostate cancer: results from the population-based PCBaSe. Lancet Oncol. 2010;11:450-458 Crossref.
  • [14] P. Vestergaard, L. Rejnmark, L. Mosekilde. Fracture risk in patients with different types of cancer. Acta Oncol. 2009;48:105-115 Crossref.
  • [15] P.W. Dickman, J. Adolfsson, K. Astrom, G. Steineck. Hip fractures in men with prostate cancer treated with orchiectomy. J Urol. 2004;172:2208-2212 Crossref.
  • [16] V.B. Shahinian, Y.F. Kuo, J.L. Freeman, J.S. Goodwin. Risk of fracture after androgen deprivation for prostate cancer. N Engl J Med. 2005;352:154-164 Crossref.
  • [17] J.A. Heit, M.D. Silverstein, D.N. Mohr, T.M. Petterson, W.M. O’Fallon, L.J. Melton III. Risk factors for deep vein thrombosis and pulmonary embolism: a population-based case-control study. Arch Intern Med. 2000;160:809-815 Crossref.
  • [18] J. Beyer, S. Wessela, O.W. Hakenberg, et al. Incidence, risk profile and morphological pattern of venous thromboembolism after prostate cancer surgery. J Thromb Haemost. 2009;7:597-604 Crossref.
  • [19] J.B. Eifler, A.W. Levinson, M.E. Hyndman, B.J. Trock, C.P. Pavlovich. Pelvic lymph node dissection is associated with symptomatic venous thromboembolism risk during laparoscopic radical prostatectomy. J Urol. 2011;185:1661-1665
  • [20] L. Barlow, K. Westergren, L. Holmberg, M. Talback. The completeness of the Swedish Cancer Register: a sample survey for year 1998. Acta Oncol. 2009;48:27-33 Crossref.
  • [21] J. Adolfsson, H. Garmo, E. Varenhorst, et al. Clinical characteristics and primary treatment of prostate cancer in Sweden between 1996 and 2005. Scand J Urol Nephrol. 2007;41:456-477 Crossref.
  • [22] Statistics in the Areas of Health and Medical Care database. 2007. http://www.socialstyrelsen.se/en/Statistics/Statistical_databases.htm.
  • [23] Central Bureau for Statistics. Statistics Sweden. Stockholm, Sweden 2008. http://www.scb.se/.
  • [24] N. Hammar, L. Alfredsson, M. Rosen, C.L. Spetz, T. Kahan, A.S. Ysberg. A national record linkage to study acute myocardial infarction incidence and case fatality in Sweden. Int J Epidemiol. 2001;30(Suppl 1):S30-S34 Crossref.
  • [25] E. Ingelsson, J. Arnlov, J. Sundstrom, L. Lind. The validity of a diagnosis of heart failure in a hospital discharge register. Eur J Heart Fail. 2005;7:787-791 Crossref.
  • [26] U. Lindblad, L. Rastam, J. Ranstam, M. Peterson. Validity of register data on acute myocardial infarction and acute stroke: the Skaraborg Hypertension Project. Scand J Soc Med. 1993;21:3-9 Crossref.
  • [27] E. Hagel, H. Garmo, A. Bill-Axelson, et al. PCBaSe Sweden: a register-based resource for prostate cancer research. Scand J Urol Nephrol. 2009;43:342-349 Crossref.
  • [28] P. Stattin, E. Holmberg, O. Bratt, J. Adolfsson, J.E. Johansson, J. Hugosson. Surveillance and deferred treatment for localized prostate cancer. Population based study in the National Prostate Cancer Register of Sweden. J Urol. 2008;180:2423-2429 discussion 2429–30
  • [29] O.S. Miettinen. The need for randomization in the study of intended effects. Stat Med. 1983;2:267-271 Crossref.
  • [30] A.V. D’Amico, J.W. Denham, J. Crook, et al. Influence of androgen suppression therapy for prostate cancer on the frequency and timing of fatal myocardial infarctions. J Clin Oncol. 2007;25:2420-2425 Crossref.
  • [31] R.D. Jones, J.E. Nettleship, D. Kapoor, H.T. Jones, K.S. Channer. Testosterone and atherosclerosis in aging men: purported association and clinical implications. Am J Cardiovasc Drugs. 2005;5:141-154 Crossref.
  • [32] B.C. Jessie, F.F. Marshall. Pharmacological prophylaxis of venous thromboembolism in contemporary radical retropubic prostatectomy: does concomitant pelvic lymphadenectomy matter?. Int J Urol. 2008;15:951-956 Crossref.
  • [33] P. Stattin, E. Holmberg, J. Johansson, L. Holmberg, J. Adolfsson, J. Hugosson. Outcomes for men with localized prostate cancer in the National Prostate Cancer Register (NPCR) of Sweden. J Natl Cancer Inst. 2010;102:950-958 Crossref.

Footnotes

a King's College London, School of Medicine, Division of Cancer Studies, Cancer Epidemiology Group, London, UK

b Regional Oncologic Centre, Uppsala University, Uppsala, Sweden

c Department of Surgical and Perioperative Sciences, Urology and Andrology, Umeå University, Umeå, Sweden

d Oncological Centre, CLINTEC Department, Karolinska Institutet, Stockholm, Sweden

lowast Corresponding author. King's College London, School of Medicine, Division of Cancer Studies, Cancer Epidemiology Group, Research Oncology, 3rd Floor, Bermondsey Wing, Guy's Hospital, London SE1 9RT, UK. Tel. +44 0 20 7188 8414.