Background:

The association of blood pressure (BP) with prostate cancer risk after accounting for asymptomatic prostate-specific antigen (PSA) testing, and with prostate cancer death, is unclear.

Methods:

We investigated BP, measured at a mean age of 38 years among 430,472 men from five Swedish cohorts, in association with incident prostate cancer (n = 32,720) and prostate cancer death (n = 6718). HRs were calculated from multivariable Cox regression models.

Results:

Increasing systolic and diastolic BP levels combined were associated with a slightly lower prostate cancer risk, with a HR of 0.98 (95% CI, 0.97–0.99) per standard deviation (SD) of mid-BP (average of systolic and diastolic BP). The association was restricted to the PSA era (1997 onwards, HR, 0.96; 95% CI, 0.95–0.98), to diagnoses initiated by a PSA test in asymptomatic men (HR, 0.95; 95% CI, 0.93–0.97), and to low-risk prostate cancer (HR, 0.95; 95% CI, 0.92–0.97). There was no clear association with more advanced disease at diagnosis. In cases, a slightly higher risk of prostate cancer death was observed for higher BP levels (HR, 1.05; 95% CI, 1.01–1.08) per SD of mid-BP; however, the association was restricted to distant metastatic disease (Pheterogeneity between case groups = 0.01), and there was no association for BP measured less than 10 years prior to diagnosis.

Conclusions:

Prediagnostic BP is unlikely an important risk factor for prostate cancer development and death. Less asymptomatic PSA testing among men with higher BP levels may explain their lower risk of prostate cancer.

Impact:

Elevated BP is unlikely to be an important risk factor for prostate cancer.

Hypertension, defined as a systolic blood pressure (BP) of 140 mmHg or higher or a diastolic BP of 90 mmHg or higher, causes a large societal disease burden with a prevalence of more than 30% worldwide in 2019 (1). Hypertension has been associated with an increased risk of several cancers (2), but its association with prostate cancer is unclear. BP increases following activation of the renin-angiotensin system (RAS) with angiotensin II as the key effector hormone, which has shown the potential to promote prostate cancer progression in several ways such as through increased oxidative stress (3–6). Oxidative stress through its reactive oxygen species may, together with the associated systemic inflammation, promote prostate cancer cell growth (7–9). Whether systolic or diastolic BP plays a greater role in these putative mechanisms is unknown.

Observational studies of hypertension and prostate cancer risk have shown heterogenous results and no association overall in meta-analyses of prospective studies (10, 11). However, few studies have investigated BP across its full exposure range. The largest studies to date are our previous studies, originating from a European population with around 6,600 incident prostate cancer cases during follow-up (2, 12), and of the Swedish Construction Workers Cohort that included 10,000 incident cases at the time (13). The latter study remains one of few studies that investigated BP levels and prostate cancer risk by low- and high-risk disease, which is essential given the large heterogeneity of prostate cancer. The study showed an inverse association with any prostate cancer and low-risk disease, and no clear association with high-risk disease, which could be clarified by finer categorization of prostate cancer, and by investigating the potential influence of an increased use of prostate-specific antigen (PSA) testing as a screening tool to detect prostate cancer.

In relation to death from prostate cancer, our aforementioned studies showed no association (13) or a positive (2, 12) association with prediagnostic BP levels. In absence of all relevant information around the time of prostate cancer diagnosis, we analyzed the associations only in the full cancer-free cohort at baseline, which, however, results in the combined association of BP with incident prostate cancer and survival after diagnosis (14). Studies of BP and prostate cancer death in cases only are lacking but use of antihypertensive drugs has been associated with a higher risk of prostate cancer death among prostate cancer cases (15, 16), possibly reflecting hypertension as the underlying cause.

In this study, we pooled five Swedish cohorts to investigate the association of BP levels with prostate cancer risk in an analysis that included three times the number of incident prostate cancers compared with the previous largest study (13). We investigated the association in detailed prostate cancer risk categories, within the pre-PSA and PSA testing era, and according to the diagnostic work-up leading to the diagnosis (PSA test or other). Associations with death from prostate cancer were investigated in the full cohort as well as in cases only incorporating detailed clinical information.

Study population

The study included men in five Swedish cohorts: the Construction Workers Cohort (“Bygghälsan”; ref. 17), the Västerbotten Intervention Programme (18, 19), the Northern Sweden Monica study (18, 20), the Malmö Diet and Cancer Study (21), and the Malmö Preventive Project (22). BP was measured as part of one or more health examinations performed in 1971–2016 (Table 1). A single BP reading was taken, except in the Malmö-based cohorts where an average of two BP readings was recorded. BP was measured in supine position using a standard mercury sphygmomanometer in all studies. In the Västerbotten Intervention Programme, BP was measured in seated position from Sept 1, 2009, onwards. These values were corrected using age- and sex-specific formula based on individuals with BP taken using both measurement methods (Supplementary Table S1). The health examination in each cohort also included measurement of height and weight, in this study incorporated as body mass index (BMI, kg/m2), and smoking status assessed from self-report questionnaires, except in the Construction Workers Cohort from 1978 onwards when smoking was determined from a standardised oral questionnaire filled out by a nurse. The study was approved by the ethics committee at Lund University (Lund, Sweden; no. 2016/564).

Table 1.

Baseline characteristics of the 430,472 men in the study, in total and across BP level.

Systolic/diastolic BP level, mmHg
Baseline characteristicsTotal (n = 430,472)<130/85 (n = 154,843)130/85–140/90 (n = 114,513)140/90–160/100 (n = 125,401)>160/100 (n = 35,715)
Cohort (examination years), n (%) 
 CWC (1971–1993) 344,602 (80) 117,339 (76) 96,880 (84) 103,983 (83) 26,400 (74) 
 VIP (1985–2016) 55,362 (13) 28,750 (19) 11,413 (10) 11,278 (9) 3,921 (11) 
 MONICA (1986–2014) 4,009 (1) 1,665 (1) 826 (1) 1,047 (1) 471 (1) 
 MDCS (1991–1996) 11,611 (3) 1,994 (1) 2,061 (2) 4,391 (3) 3,165 (9) 
 MPP (1974–1984) 14,888 (3) 5,095 (4) 3,333 (3) 4,702 (4) 1,758 (5) 
Age at study entry, y 
 Mean (SD) 37.5 (13.6) 33.4 (11.8) 34.5 (12.5) 41.1 (13.7) 52.6 (10.7) 
 Categories, n (%) 
  <35 208,724 (49) 91,700 (59) 66,940 (58) 47,256 (38) 2,828 (8) 
  35–44 87,228 (20) 35,374 (23) 22,238 (19) 25,185 (20) 4,431 (12) 
  45–54 73,800 (17) 18,941 (12) 15,609 (14) 28,263 (22) 10,987 (31) 
  ≥55 60,720 (14) 8828 (6) 9,726 (8) 24,697 (20) 17,469 (49) 
Body mass index, kg/m2 
 Mean (SD) 24.6 (3.4) 23.6 (3.0) 24.3 (3.1) 25.4 (3.4) 26.9 (3.7) 
 Categories, n (%) 
  <25 257,174 (60) 111,021 (72) 73,029 (64) 62,030 (49) 11,094 (31) 
  25–29.9 145,526 (34) 38,985 (25) 36,193 (31) 52,221 (42) 18,127 (51) 
  ≥30 27,772 (6) 4,837 (3) 5,291 (5) 11,150 (9) 6,494 (18) 
Smoking, n (%) 
 Never smoker 168,279 (39) 68,648 (44) 45,339 (40) 43,463 (35) 10,829 (30) 
 Former smoker 69,871 (16) 21,548 (14) 17,534 (15) 22,833 (18) 7,956 (22) 
 Current smoker 132,666 (31) 46,595 (30) 35,841 (31) 39,889 (32) 10,341 (29) 
 Missing 59,656 (14) 18,052 (12) 15,799 (14) 19,216 (15) 6,589 (19) 
Education, n (%)a 
 Pre-upper secondary school <9 years 112,943 (26) 26,890 (17) 27,814 (24) 43,653 (35) 14,596 (41) 
 Pre-upper secondary school 9 years 38,055 (9) 15,868 (10) 11,278 (10) 9,490 (8) 1,419 (4) 
 Max. 2 years upper secondary school 159,722 (37) 68,644 (44) 46,705 (41) 38,199 (30) 6,174 (17) 
 3 years upper secondary school 46,474 (11) 18,297 (12) 12,842 (11) 12,590 (10) 2,745 (8) 
 Post-upper secondary school <3 years 24,050 (6) 11,319 (7) 6,546 (6) 5,220 (4) 965 (3) 
 Post-upper secondary school ≥3 years 15,894 (4) 8,417 (6) 3,455 (3) 3,138 (3) 884 (2) 
 Missing 33,334 (8) 5,408 (3) 5,873 (5) 13,111 (10) 8,942 (25) 
Country of birth, n (%) 
 Born in SE and both parents born in SE 384,360 (89) 135,763 (88) 101,805 (89) 113,552 (91) 33,240 (93) 
 Other 46,112 (11) 19,080 (12) 12,708 (11) 11,849 (9) 2,475 (7) 
Systolic/diastolic BP level, mmHg
Baseline characteristicsTotal (n = 430,472)<130/85 (n = 154,843)130/85–140/90 (n = 114,513)140/90–160/100 (n = 125,401)>160/100 (n = 35,715)
Cohort (examination years), n (%) 
 CWC (1971–1993) 344,602 (80) 117,339 (76) 96,880 (84) 103,983 (83) 26,400 (74) 
 VIP (1985–2016) 55,362 (13) 28,750 (19) 11,413 (10) 11,278 (9) 3,921 (11) 
 MONICA (1986–2014) 4,009 (1) 1,665 (1) 826 (1) 1,047 (1) 471 (1) 
 MDCS (1991–1996) 11,611 (3) 1,994 (1) 2,061 (2) 4,391 (3) 3,165 (9) 
 MPP (1974–1984) 14,888 (3) 5,095 (4) 3,333 (3) 4,702 (4) 1,758 (5) 
Age at study entry, y 
 Mean (SD) 37.5 (13.6) 33.4 (11.8) 34.5 (12.5) 41.1 (13.7) 52.6 (10.7) 
 Categories, n (%) 
  <35 208,724 (49) 91,700 (59) 66,940 (58) 47,256 (38) 2,828 (8) 
  35–44 87,228 (20) 35,374 (23) 22,238 (19) 25,185 (20) 4,431 (12) 
  45–54 73,800 (17) 18,941 (12) 15,609 (14) 28,263 (22) 10,987 (31) 
  ≥55 60,720 (14) 8828 (6) 9,726 (8) 24,697 (20) 17,469 (49) 
Body mass index, kg/m2 
 Mean (SD) 24.6 (3.4) 23.6 (3.0) 24.3 (3.1) 25.4 (3.4) 26.9 (3.7) 
 Categories, n (%) 
  <25 257,174 (60) 111,021 (72) 73,029 (64) 62,030 (49) 11,094 (31) 
  25–29.9 145,526 (34) 38,985 (25) 36,193 (31) 52,221 (42) 18,127 (51) 
  ≥30 27,772 (6) 4,837 (3) 5,291 (5) 11,150 (9) 6,494 (18) 
Smoking, n (%) 
 Never smoker 168,279 (39) 68,648 (44) 45,339 (40) 43,463 (35) 10,829 (30) 
 Former smoker 69,871 (16) 21,548 (14) 17,534 (15) 22,833 (18) 7,956 (22) 
 Current smoker 132,666 (31) 46,595 (30) 35,841 (31) 39,889 (32) 10,341 (29) 
 Missing 59,656 (14) 18,052 (12) 15,799 (14) 19,216 (15) 6,589 (19) 
Education, n (%)a 
 Pre-upper secondary school <9 years 112,943 (26) 26,890 (17) 27,814 (24) 43,653 (35) 14,596 (41) 
 Pre-upper secondary school 9 years 38,055 (9) 15,868 (10) 11,278 (10) 9,490 (8) 1,419 (4) 
 Max. 2 years upper secondary school 159,722 (37) 68,644 (44) 46,705 (41) 38,199 (30) 6,174 (17) 
 3 years upper secondary school 46,474 (11) 18,297 (12) 12,842 (11) 12,590 (10) 2,745 (8) 
 Post-upper secondary school <3 years 24,050 (6) 11,319 (7) 6,546 (6) 5,220 (4) 965 (3) 
 Post-upper secondary school ≥3 years 15,894 (4) 8,417 (6) 3,455 (3) 3,138 (3) 884 (2) 
 Missing 33,334 (8) 5,408 (3) 5,873 (5) 13,111 (10) 8,942 (25) 
Country of birth, n (%) 
 Born in SE and both parents born in SE 384,360 (89) 135,763 (88) 101,805 (89) 113,552 (91) 33,240 (93) 
 Other 46,112 (11) 19,080 (12) 12,708 (11) 11,849 (9) 2,475 (7) 

Abbreviations: CWC, Construction Workers Cohort; MDCS, Malmö Diet and Cancer Study; MONICA, Northern Sweden MONICA study; MPP, Malmö Preventive Project; SE, Sweden; VIP, Västerbotten Intervention Programme.

aDetermined by the Swedish Longitudinal integration database for health insurance and labor market studies.

Follow-up

The unique personal identification number of each inhabitant of Sweden was used to follow up men in the cohorts in national registers until December 31, 2016. The Swedish Cancer Register (23) was used to capture diagnoses of prostate cancer (ICD-7 code 177) and other cancers, the Cause of Death Register (24) was used to obtain information on the underlying cause of death, and the Total Population Register was used for information on emigration. We also linked individuals to the Longitudinal Integration Database for Health Insurance and Labour Market Studies for information on sociodemographic factors (25), and to the Hospital Discharge Register for information on in-patient care (26), which we used for the Charlson Comorbidity Index (27).

Prostate cancer cases

The National Prostate Cancer Register (NPCR) of Sweden covers 99% of prostate cancer diagnoses since 1998 as compared with the Cancer Register and was used to obtain information on tumour characteristics at the time of diagnosis, and on primary treatment (28). We classified prostate cancer into five risk categories (28): localized low-risk (T1–2, Gleason score 2–6 and PSA level <10 ng/mL), localized intermediate-risk (T1–2, Gleason score 7 and/or PSA 10 to <20 ng/mL), localized high-risk (T3 and/or Gleason score 8–10 and/or PSA 20 to <50 ng/mL), regionally metastatic/locally advanced [T4 and/or N1 and/or PSA 50 to <100 ng/mL in the absence of distant metastases (M0 or Mx)], or distant metastases (M1 and/or PSA ≥100 ng/mL). We also grouped prostate cancers into “all localized” (the three lowest prostate cancer risk categories), and “all advanced” (upper two risk categories). We further analyzed our results based on the diagnostic work-up leading to the prostate cancer diagnosis, which had been recorded in the NPCR since the year 2000 as detection due to an asymptomatic PSA test at a health examination, lower urinary tract symptoms (LUTS) or other symptoms. These data have lower validity than other variables in the NPCR (29), but are not expected to be related to BP level.

Inclusion criteria

Out of 467,447 men with 1,268,174 health examinations, we excluded 67,068 examinations in men recorded in several cohorts (i.e., keeping information from one cohort), 22,231 examinations performed before 18 years of age, 4,126 examinations with a prevalent cancer, 59 examinations with mismatching dates, and 12,518 examinations with BP or BMI missing. Of the remaining 430,472 men with 1,162,172 health examinations, the first was selected as baseline observation in the analysis.

Statistical analysis

In the full-cohort analysis, person-years were calculated from the baseline date until the date of prostate cancer diagnosis, another cancer diagnosis, emigration, death, or end of follow-up, whichever came first. In analyses of a specific prostate cancer risk category as event, other prostate cancer risk categories were censored at the time of their diagnosis. Cox regression was used with attained age as time scale to estimate HRs and 95% confidence intervals (CI) of prostate cancer and prostate cancer death according to categories of systolic and diastolic BP in single and combined, and per SD increment of mid-BP (average of systolic and diastolic BP), systolic BP, and diastolic BP. Mid-BP is a strong predictor of cardiovascular mortality (30) and is reported in our study alongside the results for categories of systolic and diastolic BP combined. These were defined according to the European Society of Cardiology and the European Society of Hypertension: systolic/diastolic BP <120/80 (optimal), 120–129/80–84 (normal), 130–139/85–89 (high normal), 140–159/90–99 (grade 1 hypertension), and ≥160/100 (grade 2–3 hypertension; ref. 31). Optimal and normal BP were combined into a referent group. We stratified the Cox models by cohort and birth year (<1935/1935–39/1940–44/1945–49/≥1950), and adjusted for baseline age (continuous), smoking status (never smoking/former smoking/current smoking/missing), BMI (<25/25.0–29.9/≥30 kg/m2), education (see categories in Table 1), six geographical regions, and country of birth (born in Sweden with both parents born in Sweden/born in Sweden with one parent born in Sweden/born in Sweden with both parents born abroad/born abroad). In sensitivity analyses of incident prostate cancer, we analyzed the association with BP in subgroups of baseline age and attained age to ensure that any potential influence of age on the results had not gone undetected in our primary analyses. We also investigated the association of BP with prostate cancer risk (i) before 1997 versus from 1 Jan. 1997 onwards (proxy for the start of opportunistic PSA testing in Sweden), and (ii) for prostate cancers detected through an asymptomatic PSA test versus through LUTS or other symptoms.

In the case-only analysis of BP and prostate cancer death, we used Cox regression with time since diagnosis as the time scale and used the same strata and adjustments for baseline age, smoking, BMI, region, and country of birth as in the full-cohort analysis. We additionally adjusted for time between baseline measurement and prostate cancer diagnosis (continuous), and for variables at the time of diagnosis: age (continuous), highest level of education, civil status (unmarried/married/divorced/widower/missing), income (<158/158–193/193–230/>230 kSEK/year/missing), source of income (work/studies/care of child or family/sick/unemployed/ early retirement/social benefits/labor market policy activity/pensioner/no income/missing), Charlson Comorbidity Index (none/mild/severe; ref. 27), primary treatment (conservative/curative/noncurative/missing), and prostate cancer risk category (in the total analysis). In response to a positive association between BP and prostate cancer death in distant metastatic cases, we performed further post hoc analyses. (i) We investigated whether the adjustments for age and comorbidity at diagnosis—two potential mediators in the association—had an influence on the results, and we analyzed younger men (<70 years at diagnosis) and men without comorbidities separately in an attempt to isolate the association of prostate cancer with BP from the association with age and comorbidity. (ii) For case groups with a larger number of prostate cancer deaths including distant metastatic cases, we also analyzed the association of mid-BP with prostate cancer death among individuals with a BP measurement taken less than 10 years prior to diagnosis.

We tested the interaction between BP and cohort using the likelihood ratio test and found no indication of interaction, which supported the pooling of our data. Evaluation of the proportional hazards assumption with Schoenfeld residuals, and model stratification in case of indications for violation, showed no need for additional strata in our model. The heterogeneity between prostate cancer risk categories was calculated using the Lunn and McNeil method (32). All statistical tests were two-sided and statistical analyses were performed using STATA version 17 (StataCorp LP).

Data availability statement

The data that support the findings of this study are available from cohort committees and national registers. Restrictions apply to the availability of these data, which were used under license for this study. Data are available after contact with the corresponding author conditional on permission from the involved cohort committees and national registers.

Baseline characteristics of the 430,570 men in the study are shown in Table 1, in total, and across BP categories. The mean baseline age was 37.5 (SD = 13.6) years. For each higher BP category, baseline age and BMI were higher, and the proportions of never-smokers and men with higher education were lower. After on average 27.9 (SD = 11.7) years of follow-up, totaling 12,022,782 person-years, 32,732 men had been diagnosed with prostate cancer, of which 26,995 had clinical characteristics data available used for prostate cancer risk categorization (Supplementary Table S2). Among cases, higher baseline BP level was associated with a higher age at diagnosis and more comorbidity. In total, 6,722 cases died from prostate cancer during follow-up.

Table 2.

HR of prostate cancer according to BP level in the pre-PSA era and the PSA era, and by detection mode in the PSA era.

Pre-PSA era (1974–1996), n = 369,360aPSA era (1997–2016), n = 424,650a
BP leveln casesHR (95% CI)bn casesHR (95% CI)b
SBP/DBP, mmHg 
 <130/85 883 Reference 8,757 Reference 
 130/85–140/90 1,052 0.96 (0.88–1.05) 7,338 0.96 (0.93–0.99) 
 140/90–160/100 2,450 0.96 (0.88–1.03) 8,762 0.95 (0.92–0.98) 
 >160/100 1,437 1.00 (0.91–1.09) 2,041 0.88 (0.83–0.92) 
Ptrend  1.0  <0.001 
Mid-BP, per SD 5,822 1.00 (0.98–1.03) 26,898 0.96 (0.95–0.98) 
   PSA era (2000–2016), n = 355,666a 
   Asymptomatic, PSA-detected Detection through LUTS or other symptoms 
   n cases HR (95% CI)b n cases HR (95% CI)b 
SBP/DBP, mmHg 
 <130/85   3,580 Reference 4,013 Reference 
 130/85–140/90   2,793 0.97 (0.92–1.01) 3,475 0.98 (0.94–1.03) 
 140/90–160/100   2,674 0.93 (0.88–0.97) 4,523 1.01 (0.96–1.05) 
 >160/100   430 0.83 (0.74–0.92) 1,112 1.02 (0.95–1.09) 
Ptrend    <0.001  0.6 
Mid-BP, per SD   9,477 0.95 (0.93–0.97) 13,123 1.01 (0.99–1.02)c 
Pre-PSA era (1974–1996), n = 369,360aPSA era (1997–2016), n = 424,650a
BP leveln casesHR (95% CI)bn casesHR (95% CI)b
SBP/DBP, mmHg 
 <130/85 883 Reference 8,757 Reference 
 130/85–140/90 1,052 0.96 (0.88–1.05) 7,338 0.96 (0.93–0.99) 
 140/90–160/100 2,450 0.96 (0.88–1.03) 8,762 0.95 (0.92–0.98) 
 >160/100 1,437 1.00 (0.91–1.09) 2,041 0.88 (0.83–0.92) 
Ptrend  1.0  <0.001 
Mid-BP, per SD 5,822 1.00 (0.98–1.03) 26,898 0.96 (0.95–0.98) 
   PSA era (2000–2016), n = 355,666a 
   Asymptomatic, PSA-detected Detection through LUTS or other symptoms 
   n cases HR (95% CI)b n cases HR (95% CI)b 
SBP/DBP, mmHg 
 <130/85   3,580 Reference 4,013 Reference 
 130/85–140/90   2,793 0.97 (0.92–1.01) 3,475 0.98 (0.94–1.03) 
 140/90–160/100   2,674 0.93 (0.88–0.97) 4,523 1.01 (0.96–1.05) 
 >160/100   430 0.83 (0.74–0.92) 1,112 1.02 (0.95–1.09) 
Ptrend    <0.001  0.6 
Mid-BP, per SD   9,477 0.95 (0.93–0.97) 13,123 1.01 (0.99–1.02)c 

Abbreviations: DBP, diastolic BP; SBP, systolic BP.

aThe analysis of the pre-PSA era included men with a study entry before January 1, 1997 and the follow-up in the analysis ended on December 31, 1996. The analysis of the PSA-era included men on study as of January 1, 1997, and the detection mode analysis included all men on study as of January 1, 2000.

bHRs were derived from Cox regression models with attained age as time scale, stratified on cohort and birthyear, and adjusted for baseline age, body mass index, smoking status, region, country of birth, and education. The P values for trend in HRs across BP categories were derived from the Wald test of integer scores of BP categories.

cThe HR for LUTS-detected (0.99; 95% CI, 0.96–1.02) versus other symptoms-detected (1.00; 95% CI, 0.96–1.01) prostate cancer did not differ. Separate information for LUTS versus other symptoms was only available from 2004 onwards.

Elevated BP was associated with a small, lower risk of any prostate cancer; the HR was 0.96 (95% CI, 0.92–0.99) for systolic/diastolic BP ≥160/100 versus <130/85 (Fig. 1), and 0.98 (95% CI, 0.97–0.99) per SD of mid-BP (Fig. 2). This inverse association was only evident in the PSA era, and within the PSA era only for prostate cancers detected through a PSA test and not through LUTS or other symptoms (Table 2). The association between BP and prostate cancer risk changed across prostate cancer risk categories (Pheterogeneity = 0.02), ranging from a weak inverse association for low-risk prostate cancer (HR per SD mid-BP, 0.95; 95% CI, 0.92–0.97) to a weak nonsignificant positive association for distant metastatic disease (HR per SD mid-BP, 1.04; 95% CI, 1.00–1.08). There was no association of BP with prostate cancer death in the full-cohort analysis of all men followed from baseline (Figs. 1 and 2).

Figure 1.

HR (95% CI) of any incident prostate cancer and by prostate cancer risk category, and of prostate cancer death, according to BP level among the 430,472 men in the study. The analysis by prostate cancer risk category included 365,875 noncensored men by January 1, 1998 when the National Prostate Cancer Register of Sweden became nationwide. HRs were derived from Cox regression models with attained age as time scale, stratified on cohort and birthyear, and adjusted for baseline age, body mass index, smoking status, region, country of birth, and education. The P values for trend in HRs across BP categories were derived from the Wald test of BP as integer scores. The P value for the heterogeneity in HRs between prostate cancer risk categories was calculated using the Lunn and McNeil duplication method.

Figure 1.

HR (95% CI) of any incident prostate cancer and by prostate cancer risk category, and of prostate cancer death, according to BP level among the 430,472 men in the study. The analysis by prostate cancer risk category included 365,875 noncensored men by January 1, 1998 when the National Prostate Cancer Register of Sweden became nationwide. HRs were derived from Cox regression models with attained age as time scale, stratified on cohort and birthyear, and adjusted for baseline age, body mass index, smoking status, region, country of birth, and education. The P values for trend in HRs across BP categories were derived from the Wald test of BP as integer scores. The P value for the heterogeneity in HRs between prostate cancer risk categories was calculated using the Lunn and McNeil duplication method.

Close modal
Figure 2.

HR (95% CI) of any incident prostate cancer and by prostate cancer risk category, and of prostate cancer death, per SD of mid-BP level among the 430,472 men in the study. The analysis by prostate cancer risk category included 365,875 noncensored men by January 1, 1998 when the National Prostate Cancer Register of Sweden became nationwide. See Fig. 1 for information on the statistical model.

Figure 2.

HR (95% CI) of any incident prostate cancer and by prostate cancer risk category, and of prostate cancer death, per SD of mid-BP level among the 430,472 men in the study. The analysis by prostate cancer risk category included 365,875 noncensored men by January 1, 1998 when the National Prostate Cancer Register of Sweden became nationwide. See Fig. 1 for information on the statistical model.

Close modal

In survival analysis of prostate cancer cases adjusted for clinical characteristics, elevated BP levels were associated with a higher risk of prostate cancer death; the HR was 1.18 (95% CI, 1.05–1.32) for systolic/diastolic BP ≥160/100 versus <130/85 (Fig. 3), and 1.05 (95% CI, 1.01–1.08) per SD of mid-BP (Fig. 4). The results differed by prostate cancer risk category at diagnosis (Pheterogeneity = 0.01) and was positive only for distant metastatic prostate cancer (HR per SD mid-BP, 1.08; 95% CI, 1.02–1.13). In distant metastatic cases, the adjustments for age and comorbidity at the time of diagnosis had no effect on the HR (identical HRs per SD mid-BP), and the positive association with BP remained when restricting the analysis to men with a Charlson Comorbidity Index of null (HR per SD mid-BP, 1.06 (95% CI, 1.00–1.12) or to men younger than 70 years at diagnosis (HR per SD mid-BP, 1.11; 95% CI, 1.02–1.22). Among cases with a BP measurement less than 10 years prior to the prostate cancer diagnosis, there was no association with prostate cancer death overall or in distant metastatic cases (Fig. 4).

Figure 3.

HR (95% CI) of prostate cancer death according to BP level amongst the 26,995 prostate cancer cases in the study. HRs were derived from Cox regression models with time since diagnosis as time scale, stratified on cohort and birthyear, and adjusted for baseline body mass index, smoking status, region, country of birth, time between baseline and diagnosis, and for variables at the time of diagnosis: age, civil status, education, income, source of income, Charlson Comorbidity Index, primary treatment, and prostate cancer risk category in the all prostate cancers analysis. The P values for trend in hazard ratios across BP categories were derived from the Wald test of BP as integer scores. The P value for the heterogeneity in HRs between prostate cancer risk categories was calculated using the Lunn and McNeil duplication method.

Figure 3.

HR (95% CI) of prostate cancer death according to BP level amongst the 26,995 prostate cancer cases in the study. HRs were derived from Cox regression models with time since diagnosis as time scale, stratified on cohort and birthyear, and adjusted for baseline body mass index, smoking status, region, country of birth, time between baseline and diagnosis, and for variables at the time of diagnosis: age, civil status, education, income, source of income, Charlson Comorbidity Index, primary treatment, and prostate cancer risk category in the all prostate cancers analysis. The P values for trend in hazard ratios across BP categories were derived from the Wald test of BP as integer scores. The P value for the heterogeneity in HRs between prostate cancer risk categories was calculated using the Lunn and McNeil duplication method.

Close modal
Figure 4.

HR (95% CI) of prostate cancer death per SD of mid-BP level among the 26,995 prostate cancer cases in the study, and amongst cases with a BP measurement less than 10 years prior to the diagnosis. See Fig. 3 for information on the statistical model.

Figure 4.

HR (95% CI) of prostate cancer death per SD of mid-BP level among the 26,995 prostate cancer cases in the study, and amongst cases with a BP measurement less than 10 years prior to the diagnosis. See Fig. 3 for information on the statistical model.

Close modal

Separate results for systolic and diastolic BP and prostate cancer risk and death are shown in Supplementary Tables S3 and S4. There were consistently overlapping CIs between the two BP indices per SD increment, except for total incident prostate cancer, which showed an inverse association only for systolic BP.

Analyses of BP and risk of total, localized, and advanced prostate cancer showed no differences in associations in subgroups of baseline or attained age, except for a stronger inverse association with total prostate cancer in men with an attained age of less than 60 years than in older men (Supplementary Table S5). Cases in this younger age-group were diagnosed during later years (median, 2010 vs. 2000 for >70 years), and were more often diagnosed with localized low-risk prostate cancer (47% vs. 15% for >70 years).

This large, pooled, Swedish cohort study showed a slightly lower risk of prostate cancer in men with elevated BP levels. The association was restricted to low-risk prostate cancer, prostate cancer diagnosed in the “PSA-era” (1997 onwards), and within this era, to prostate cancers detected through an asymptomatic PSA test. Elevated BP prior to prostate cancer diagnosis was associated with poorer survival from prostate cancer, but the results were heterogeneous between case groups and were mainly driven by distant metastatic prostate cancer, and there was no association among cases with a BP measurement taken closer than 10 years before diagnosis.

In a previous study of the Swedish Construction Workers Cohort, which made up approximately 80% of the men in the present study but had 12 years shorter follow-up time, we found a lower risk of any and low-risk prostate cancer in men with elevated BP levels (13). In this study, we found that these inverse associations were confined to the most indolent prostate cancers and to non-organised PSA testing that started around the mid-1990s in Sweden and has persisted since (33, 34). A similar time-trend for the BP-prostate cancer association was found in our previous study of a European population (12). Furthermore, this study showed that in the PSA era the lower risk was confined to prostate cancer recorded as detected through a PSA test taken as part of a general health check-up. These results suggest more PSA testing in men with lower BP levels. In the UK Biobank, however, men who had taken at least one PSA test were more often hypertensive, but were also more often normal weight, nonsmoking and had a healthy diet (35). This contradiction might be explained by that more health-conscious men, reflected by the latter factors, actively undertake asymptomatic PSA testing more, whereas hypertensive men may more often take a PSA test as part of the diagnostic work-up of the hypertension or a related condition. In support of this theory, men with type 2 diabetes do less PSA testing (35, 36), but more if they are on metformin, commonly prescribed in early diabetes stages, than if they are treated with insulin (36). We suggest that a higher uptake of PSA testing in men with low BP is the most likely explanation for their higher risk of prostate cancer, in particular low-risk prostate cancer.

Other larger prospective studies of BP and prostate cancer risk than the aforementioned studies are lacking, but more studies exist on the association between hypertension, commonly including the use of antihypertensive use in the definition, and prostate cancer risk (10, 11). The most recent meta-analyses of hypertension and prostate cancer risk showed inconsistent results between studies (10, 11), and no association in prospective studies totaling 22,000 incident prostate cancer cases (10), or in one later large prospective study (37). An association between hypertension and prostate cancer risk could be mediated by antihypertensive drugs; however, randomized (38) and observational (39) studies of antihypertensive use and prostate cancer risk have shown no association overall. A recent, large observational study showed a higher risk of prostate cancer, especially metastatic disease, for men on antihypertensive drugs, but the associations diminished when excluding the first years of follow-up (40). Altogether, the accumulated evidence to date suggests that elevated BP and use of antihypertensive drugs are not causal risk factors for prostate cancer.

In contrast to prostate cancer risk, we found that elevated BP was associated with a higher risk of prostate cancer death. The weaker association in the analysis of the full cohort is influenced by the inverse, albeit weak, association between BP and incident prostate cancer, inherent to the follow-up from baseline in this analysis (14). In support of hypertension potentially promoting prostate cancer progression, small patient studies have provided some (41, 42), but inconsistent (43), evidence for higher recurrence after radical prostatectomy in hypertensive men, and register studies have suggested poorer survival from prostate cancer in men on antihypertensives (15, 16). Activation of the RAS (3–6), oxidative stress, and systemic inflammation (7–9), are some of the potential mechanisms associated with hypertension that could promote prostate cancer progression. However, our heterogenous results between case groups, with an association between BP and prostate cancer death only in distant metastatic prostate cancer, speak against a causal effect of BP on prostate cancer death. Moreover, there was no association in analyses restricted to BP measurements closer to diagnosis. An association only for elevated BP long before prostate cancer diagnosis may not necessarily be less causal, as it could result from a chain of reactions promoting tumor progression after tumor initiation. However, our heterogenous findings between case groups and by lag-time with unclear explanations, question the causal and clinical relevance of BP level on prostate cancer–specific death.

Strengths of this study include the large sample size with long follow-up, detailed and validated prostate cancer clinical information (29), and information on the most important potential risk factors for prostate cancer. The prospectively measured BP level is a further strength; however, several assessments over time, especially closer to and after diagnosis, would have strengthened the investigation. A weakness of the study was that we based our calculations on a single BP reading, or an average of two readings during a single examination, which includes short-term random error due to instrumental and true day-to-day variation, but also long-term intra-individual variation of BP (44). We further lacked information on antihypertensive therapy at baseline and during follow-up, which could mediate or modify our observed associations.

In conclusion, this large, prospective study of men in Sweden showed a slightly lower risk of any and low-risk prostate cancer in men with elevated BP levels, however, with a suggested influence of detection bias, enabled by Sweden's increased opportunistic PSA testing over time and potentially more active undertaking of PSA testing in men with BP in the lower than the higher range. Elevated BP was associated with a higher risk of death from prostate cancer, but heterogenous findings across prostate cancer risk categories, and the lack of an association for BP measurements closer to diagnosis, raises doubts about the causal and clinical relevance of these findings. Prediagnostic BP level is unlikely to materially affect prostate cancer development and death.

No disclosures were reported.

S.H.J. Jochems: Conceptualization, data curation, formal analysis, investigation, writing–review and editing. C. Häggström: Investigation, writing–review and editing. P. Stattin: Data curation, investigation, writing–review and editing. B. Järvholm: Data curation, investigation, writing–review and editing. T. Stocks: Conceptualization, data curation, funding acquisition, investigation, writing–original draft.

We thank the Construction Workers Cohort (“Bygghälsan”) for providing extensive data to the study. We thank the Biobank Research Unit at Umeå University, the Västerbotten Intervention Programme, the Northern Sweden MONICA study, and the County Council of Västerbotten for providing data, and acknowledge the contribution of Biobank Sweden, supported by the Swedish Research Council (VR 2017-00650). We also thank Anders Dahlin, database manager of the MDCS and MPP cohorts, and further acknowledge the support of these cohorts by a Lund University Infrastructure grant (STYR 2019/2046). We thank the National Prostate Cancer Register of Sweden (NPCR) steering group: Pär Stattin (chairman), Ingela Franck Lissbrant (co-chair), Camilla Thellenberg, Johan Styrke, Hampus Nugin, Stefan Carlsson, David Robinson, Mats Andén, Jon Kindblom, Olof Ståhl, Tomas Jiborn, Maria Nyberg, and Fredrik Sandin. This study was supported by the Swedish Research Council (2015-02332 and 2018-02825, to T. Stocks), the Swedish Cancer Society (CAN 2017/1019, to T. Stocks), and the Cancer Research Foundation at the Department of Oncology, Malmö University Hospital, Sweden (to T. Stocks).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
NCD-Risk Collaboration
.
Worldwide trends in hypertension prevalence and progress in treatment and control from 1990 to 2019: a pooled analysis of 1201 population-representative studies with 104 million participants
.
Lancet
2021
;
398
:
957
80
.
2.
Stocks
T
,
Van Hemelrijck
M
,
Manjer
J
,
Bjorge
T
,
Ulmer
H
,
Hallmans
G
, et al
.
Blood pressure and risk of cancer incidence and mortality in the metabolic syndrome and cancer project
.
Hypertension
2012
;
59
:
802
10
.
3.
Uemura
H
,
Ishiguro
H
,
Ishiguro
Y
,
Hoshino
K
,
Takahashi
S
,
Kubota
Y
.
Angiotensin II induces oxidative stress in prostate cancer
.
Mol Cancer Res
2008
;
6
:
250
8
.
4.
Bose
SK
,
Gibson
W
,
Giri
S
,
Nath
N
,
Donald
CD
.
Angiotensin II up-regulates PAX2 oncogene expression and activity in prostate cancer via the angiotensin II type I receptor
.
Prostate
2009
;
69
:
1334
42
.
5.
Dominska
K
,
Ochedalski
T
,
Kowalska
K
,
Matysiak-Burzynska
ZE
,
Pluciennik
E
,
Piastowska-Ciesielska
AW
.
Interaction between angiotensin II and relaxin 2 in the progress of growth and spread of prostate cancer cells
.
Int J Oncol
2016
;
48
:
2619
28
.
6.
Chow
L
,
Rezmann
L
,
Catt
KJ
,
Louis
WJ
,
Frauman
AG
,
Nahmias
C
, et al
.
Role of the renin-angiotensin system in prostate cancer
.
Mol Cell Endocrinol
2009
;
302
:
219
29
.
7.
Paschos
A
,
Pandya
R
,
Duivenvoorden
WC
,
Pinthus
JH
.
Oxidative stress in prostate cancer: changing research concepts towards a novel paradigm for prevention and therapeutics
.
Prostate Cancer Prostatic Dis
2013
;
16
:
217
25
.
8.
Gupta-Elera
G
,
Garrett
AR
,
Robison
RA
,
O'Neill
KL
.
The role of oxidative stress in prostate cancer
.
Eur J Cancer Prev
2012
;
21
:
155
62
.
9.
Sfanos
KS
,
Yegnasubramanian
S
,
Nelson
WG
,
De Marzo
AM
.
The inflammatory microenvironment and microbiome in prostate cancer development
.
Nat Rev Urol
2018
;
15
:
11
24
.
10.
Liang
Z
,
Xie
B
,
Li
J
,
Wang
X
,
Wang
S
,
Meng
S
, et al
.
Hypertension and risk of prostate cancer: a systematic review and meta-analysis
.
Sci Rep
2016
;
6
:
31358
.
11.
Seretis
A
,
Cividini
S
,
Markozannes
G
,
Tseretopoulou
X
,
Lopez
DS
,
Ntzani
EE
, et al
.
Association between blood pressure and risk of cancer development: a systematic review and meta-analysis of observational studies
.
Sci Rep
2019
;
9
:
8565
.
12.
Haggstrom
C
,
Stocks
T
,
Ulmert
D
,
Bjorge
T
,
Ulmer
H
,
Hallmans
G
, et al
.
Prospective study on metabolic factors and risk of prostate cancer
.
Cancer
2012
;
118
:
6199
206
.
13.
Stocks
T
,
Hergens
MP
,
Englund
A
,
Ye
W
,
Stattin
P
.
Blood pressure, body size and prostate cancer risk in the Swedish construction workers cohort
.
Int J Cancer
2010
;
127
:
1660
8
.
14.
Cespedes Feliciano
EM
,
Prentice
RL
,
Aragaki
AK
,
Neuhouser
ML
,
Banack
HR
,
Kroenke
CH
, et al
.
Methodological considerations for disentangling a risk factor's influence on disease incidence versus postdiagnosis survival: The example of obesity and breast and colorectal cancer mortality in the women's health initiative
.
Int J Cancer
2017
;
141
:
2281
90
.
15.
Siltari
A
,
Murtola
TJ
,
Talala
K
,
Taari
K
,
Tammela
TLJ
,
Auvinen
A
.
Antihypertensive drug use and prostate cancer-specific mortality in Finnish men
.
PLoS One
2020
;
15
:
e0234269
.
16.
Santala
EE
,
Rannikko
A
,
Murtola
TJ
.
Antihypertensive drugs and prostate cancer survival after radical prostatectomy in Finland-a nationwide cohort study
.
Int J Cancer
2019
;
144
:
440
7
.
17.
Jackson
JA
,
Olsson
D
,
Punnett
L
,
Burdorf
A
,
Jarvholm
B
,
Wahlstrom
J
.
Occupational biomechanical risk factors for surgically treated ulnar nerve entrapment in a prospective study of male construction workers
.
Scand J Work Env Hea
2019
;
45
:
63
72
.
18.
Hallmans
G
,
Agren
A
,
Johansson
G
,
Johansson
A
,
Stegmayr
B
,
Jansson
JH
, et al
.
Cardiovascular disease and diabetes in the Northern Sweden Health and Disease Study Cohort - evaluation of risk factors and their interactions
.
Scand J Public Health
2003
;
31
:
18
24
.
19.
Norberg
M
,
Wall
S
,
Boman
K
,
Weinehall
L
.
The Vasterbotten Intervention Programme: background, design and implications
.
Global Health Action
2010
;
3
.
20.
Eriksson
M
,
Forslund
AS
,
Jansson
JH
,
Soderberg
S
,
Wennberg
M
,
Eliasson
M
.
Greater decreases in cholesterol levels among individuals with high cardiovascular risk than among the general population: the northern Sweden MONICA study 1994 to 2014
.
Eur Heart J
2016
;
37
:
1985
92
.
21.
Manjer
J
,
Elmstahl
S
,
Janzon
L
,
Berglund
G
.
Invitation to a population-based cohort study: differences between subjects recruited using various strategies
.
Scand J Public Health
2002
;
30
:
103
12
.
22.
Westerdahl
C
,
Zoller
B
,
Arslan
E
,
Erdine
S
,
Nilsson
PM
.
Morbidity and mortality risk among patients with screening-detected severe hypertension in the Malmo preventive project
.
J Hypertens
2014
;
32
:
2378
84
.
23.
Barlow
L
,
Westergren
K
,
Holmberg
L
,
Talback
M
.
The completeness of the Swedish cancer register - a sample survey for year 1998
.
Acta Oncol
2009
;
48
:
27
33
.
24.
Brooke
HL
,
Talback
M
,
Hornblad
J
,
Johansson
LA
,
Ludvigsson
JF
,
Druid
H
, et al
.
The Swedish cause of death register
.
Eur J Epidemiol
2017
;
32
:
765
73
.
25.
Ludvigsson
JF
,
Svedberg
P
,
Olen
O
,
Bruze
G
,
Neovius
M
.
The longitudinal integrated database for health insurance and labour market studies (LISA) and its use in medical research
.
Eur J Epidemiol
2019
;
34
:
423
37
.
26.
Ludvigsson
JF
,
Andersson
E
,
Ekbom
A
,
Feychting
M
,
Kim
JL
,
Reuterwall
C
, et al
.
External review and validation of the Swedish national inpatient register
.
BMC Public Health
2011
;
11
:
450
.
27.
Charlson
ME
,
Pompei
P
,
Ales
KL
,
MacKenzie
CR
.
A new method of classifying prognostic comorbidity in longitudinal studies: development and validation
.
J Chronic Dis
1987
;
40
:
373
83
.
28.
Van Hemelrijck
M
,
Wigertz
A
,
Sandin
F
,
Garmo
H
,
Hellstrom
K
,
Fransson
P
, et al
.
Cohort profile: The National Prostate Cancer Register of Sweden and Prostate Cancer Data Base Sweden 2.0
.
Int J Epidemiol
2013
;
42
:
956
67
.
29.
Tomic
K
,
Sandin
F
,
Wigertz
A
,
Robinson
D
,
Lambe
M
,
Stattin
P
.
Evaluation of data quality in the national prostate cancer register of Sweden
.
Eur J Cancer
2015
;
51
:
101
11
.
30.
Lewington
S
,
Clarke
R
,
Qizilbash
N
,
Peto
R
,
Collins
R
.
Age-specific relevance of usual blood pressure to vascular mortality: a meta-analysis of individual data for one million adults in 61 prospective studies
.
Lancet
2002
;
360
:
1903
13
.
31.
Williams
B
,
Mancia
G
,
Spiering
W
,
Agabiti Rosei
E
,
Azizi
M
,
Burnier
M
, et al
.
2018 ESC/ESH guidelines for the management of arterial hypertension
.
Eur Heart J
2018
;
39
:
3021
104
.
32.
Lunn
M
,
McNeil
D
.
Applying Cox regression to competing risks
.
Biometrics
1995
;
51
:
524
32
.
33.
Jonsson
H
,
Holmstrom
B
,
Duffy
SW
,
Stattin
P
.
Uptake of prostate-specific antigen testing for early prostate cancer detection in Sweden
.
Int J Cancer
2011
;
129
:
1881
8
.
34.
Enblad
AP
,
Bergengren
O
,
Andren
O
,
Larsson
A
,
Fall
K
,
Johansson
E
, et al
.
PSA testing patterns in a large Swedish cohort before the implementation of organized PSA testing
.
Scand J Urol
2020
;
54
:
376
81
.
35.
Littlejohns
TJ
,
Travis
RC
,
Key
TJ
,
Allen
NE
.
Lifestyle factors and prostate-specific antigen (PSA) testing in UK Biobank: implications for epidemiological research
.
Cancer Epidemiol
2016
;
45
:
40
46
.
36.
Beckmann
K
,
Crawley
D
,
Nordstrom
T
,
Aly
M
,
Olsson
H
,
Lantz
A
, et al
.
Association between antidiabetic medications and prostate-specific antigen levels and biopsy results
.
JAMA Netw Open
2019
;
2
:
e1914689
.
37.
Monroy-Iglesias
MJ
,
Russell
B
,
Crawley
D
,
Allen
NE
,
Travis
RC
,
Perez-Cornago
A
, et al
.
Metabolic syndrome biomarkers and prostate cancer risk in the UK Biobank
.
Int J Cancer
2021
;
148
:
825
34
.
38.
Copland
E
,
Canoy
D
,
Nazarzadeh
M
,
Bidel
Z
,
Ramakrishnan
R
,
Woodward
M
, et al
.
Antihypertensive treatment and risk of cancer: an individual participant data meta-analysis
.
Lancet Oncol
2021
;
22
:
558
70
.
39.
Cao
L
,
Zhang
S
,
Jia
CM
,
He
W
,
Wu
LT
,
Li
YQ
, et al
.
Antihypertensive drugs use and the risk of prostate cancer: a meta-analysis of 21 observational studies
.
BMC Urol
2018
;
18
:
17
.
40.
Siltari
A
,
Murtola
TJ
,
Talala
K
,
Taari
K
,
Tammela
TLJ
,
Auvinen
A
.
Antihypertensive drugs and prostate cancer risk in a Finnish population-based cohort
.
Scand J Urol
2018
;
52
:
321
7
.
41.
Asmar
R
,
Beebe-Dimmer
JL
,
Korgavkar
K
,
Keele
GR
,
Cooney
KA
.
Hypertension, obesity and prostate cancer biochemical recurrence after radical prostatectomy
.
Prostate Cancer Prostatic Dis
2013
;
16
:
62
6
.
42.
Shiota
M
,
Yokomizo
A
,
Takeuchi
A
,
Imada
K
,
Kiyoshima
K
,
Inokuchi
J
, et al
.
The feature of metabolic syndrome is a risk factor for biochemical recurrence after radical prostatectomy
.
J Surg Oncol
2014
;
110
:
476
81
.
43.
Macleod
LC
,
Chery
LJ
,
Hu
EY
,
Zeliadt
SB
,
Holt
SK
,
Lin
DW
, et al
.
Metabolic syndrome, dyslipidemia and prostate cancer recurrence after primary surgery or radiation in a veterans cohort
.
Prostate Cancer Prostatic Dis
2015
;
18
:
190
5
.
44.
Wood
AM
,
White
I
,
Thompson
SG
,
Lewington
S
,
Danesh
J
.
Regression dilution methods for meta-analysis: assessing long-term variability in plasma fibrinogen among 27,247 adults in 15 prospective studies
.
Int J Epidemiol
2006
;
35
:
1570
8
.

Supplementary data