EARLY-LIFE CARDIORESPIRATORY FITNESS AND LONG-TERM RISK OF PROSTATE CANCER

Abstract

Background:

Adolescence is a period of rapid prostatic growth, yet is understudied for susceptibility for future risk of prostate cancer (PC). We examined cardiorespiratory fitness (CRF) in late adolescence in relation to long-term PC risk.

Methods:

A population-based cohort study was conducted of all 699,125 Swedish military conscripts during 1972–1985 (97–98% of 18-year-old men) in relation to risk of PC overall, aggressive PC, and PC mortality during 1998–2017 (ages 50–65 years). CRF was measured by maximal aerobic workload, and PC was ascertained using the National Prostate Cancer Register. Muscle strength was examined as a secondary predictor.

Results:

In 38.8 million person-years of follow-up, 10,782 (1.5%) men were diagnosed with PC. Adjusting for sociodemographic factors, height, weight, and family history of PC, high CRF was associated with a slightly increased risk of any PC (highest vs. lowest quintile: incidence rate ratio [IRR], 1.10; 95% CI, 1.03–1.19; P=0.008), but was not significantly associated with aggressive PC (1.01; 0.85–1.21; P=0.90) nor PC mortality (1.24; 0.73–2.13; P=0.42). High muscle strength also was associated with a modestly increased risk of any PC (highest vs. lowest quintile: IRR, 1.14; 95% CI, 1.07–1.23; P<0.001), but not aggressive PC (0.88; 0.74–1.04; P=0.14) nor PC mortality (0.81; 0.48–1.37; P=0.43).

Conclusions:

High CRF or muscle strength in late adolescence was associated with slightly increased future risk of PC, possibly related to increased screening, but not risk of aggressive PC nor PC mortality.

Impact:

These findings illustrate the importance of distinguishing aggressive from indolent PC and assessing for potential detection bias.

INTRODUCTION

Prostate cancer (PC) is the second most common cancer in men and a leading cause of cancer mortality worldwide (1,2). Early-life exposures may increase risk of PC later in life. Autopsy studies have reported that 10–30% of men in their 30s have histological foci of PC (3–5). Adolescence is a period of rapid prostatic growth (6), during which the prostate may be more susceptible to carcinogenic exposures (7), as previously demonstrated for the breast and other maturing organs (8). Carcinogenesis models have suggested that the initiating genomic alteration for PC often occurs in adolescence (9) and develops into PC precursor lesions observed in >50% of men in the next decade of life (10). Most studies of potential PC risk factors have focused on mid- or late-life exposures, long after the prostate has developed. Identification of modifiable early-life risk factors could enable men to reduce their lifetime risk for PC, and particularly aggressive PC which has high mortality.

Cardiorespiratory fitness (CRF) is a modifiable factor that may be associated with reduced PC risk through its associations with altered immune function, inflammatory cytokines, and insulin-like growth factors (IGFs) (11–16). High CRF may enhance immune function by increasing the number and activity of natural killer cells, neutrophils, and macrophages, which are tumor-suppressive, and by reducing circulating cytokines (e.g., IL6, TNFα), which have proliferative and anti-apoptotic effects (14,15). Men with high CRF also have been reported to have lower levels of IGF-1 which promotes cell growth, and higher levels of IGF-binding proteins which have tumor-suppressor effects (11,12). Prostate biopsy studies have reported increased numbers of apoptotic PC cells in young men with high exercise levels (11–14).

Despite this evidence, CRF has seldom been examined in relation to PC risk because of the difficulty of measuring CRF in large cohorts of men and the decades of follow-up required. A few studies have examined CRF in mid-adulthood in relation to overall PC, but have yielded inconsistent findings that are difficult to interpret due to limited sample sizes and inability to distinguish aggressive from indolent PC (17–22). Other studies have examined self-reported physical activity in relation to overall PC risk, with inconclusive results (23–25). However, self-reported physical activity is a poor proxy for CRF (26,27), the underlying physiologic factor affected by physical activity that may influence cancer risks, and is difficult to recall decades later. In contrast, CRF is objectively measured as VO₂ max (maximal oxygen uptake), which measures the ability of the cardiovascular and respiratory systems to supply oxygen to skeletal muscles during sustained physical activity (28). To our knowledge, no prior studies have examined early-life CRF in relation to future risk of PC, which may help elucidate modifiable areas for PC prevention. We conducted a national cohort study of nearly 700,000 men in Sweden with objective measurement of CRF at age 18 years. Our aim was to examine early-life CRF in relation to the long-term risk of PC overall, aggressive disease, and mortality.

MATERIALS AND METHODS

Study Population

We identified 733,602 young men (mean age 18.3 ± 0.8 years) who underwent a military conscription examination in Sweden during 1972–1985. These years were chosen to coincide with the initiation of CRF testing in 1972 and to have at least 32 years of follow-up to age 50 years or older. The military conscription examination was compulsory for all 18-year-old men nationally each year except for 2–3% who either were incarcerated or had severe chronic medical conditions or disabilities documented by a physician. We excluded 34,477 (4.7%) men who had missing information for CRF during this period, leaving 699,125 men (95.3% of the original cohort) for analysis. This study was conducted in accordance with the Declaration of Helsinki and was approved by the Regional Ethics Review Board in Lund, Sweden (No. 2013/736). Participant consent was not required as this study used only de-identified registry-based secondary data.

Exposure Measurement

CRF, muscle strength, height, and weight measurements at age 18 years were obtained from the Swedish Military Conscription Registry (29–36). CRF was measured as the maximal aerobic workload in Watts using a well-validated electrically-braked stationary bicycle ergometer test, as described previously (29–37). Maximal aerobic workload is highly correlated with maximal oxygen uptake (VO₂ max; correlation ~0.9) (38), and measurement using this bicycle ergometer test is highly reproducible, with a test-retest correlation of 0.95 (39). CRF measured in this manner was examined alternatively as a continuous variable and categorical variable in quintiles (<219, 219–239, 240–260, 261–290, ≥291 Watts).

In addition to CRF, we examined muscle strength as a secondary predictor of interest because it represents a different aspect of physical fitness (29–36). Muscle strength was measured in Newtons using well-validated isometric dynamometer tests and calculated as the weighted sum of maximal knee extension, elbow flexion, and hand grip, as described previously (29–36,40). Muscle strength was examined alternatively as a continuous variable and categorical variable in quintiles (<1810, 1810–1982, 1983–2139, 2140–2324, ≥2325 Newtons).

Height and weight at age 18 years were measured using standard protocols and modeled alternatively as continuous variables and categorical variables in quintiles (29–36). Body mass index (BMI) also was examined as an alternative to height and weight. BMI was calculated as body weight in kilograms divided by the square of height in meters, and examined alternatively as a continuous and categorical variable using Centers for Disease Control and Prevention (CDC) definitions for adolescents up to age 19 years to facilitate comparability with US studies: “overweight” is defined as ≥85^th and <95^th percentile and “obesity” as ≥95^th percentile on the CDC’s 2000 sex-specific BMI-for-age growth charts, which correspond to BMI ≥25.6 and <29.0 and BMI ≥29.0, respectively, for 18-year-old males (41). In the present study, “normal BMI” refers to <85^th percentile, which corresponds to BMI <25.6 for 18-year-old males.

Prostate Cancer Ascertainment

The study cohort was followed up for the earliest diagnosis of PC from January 1, 1998, through December 31, 2017, as identified using the National Prostate Cancer Register (NPCR) of Sweden. This register contains 98% of all incident PC cases since 1998 compared with the National Cancer Registry to which reporting is mandated by law (42). NPCR also contains data on cancer characteristics including tumor grade according to Gleason score, disease stage according to the tumor, nodes, metastasis (TNM) classification, and PSA level at diagnosis.

Aggressive PC was defined by clinical stage T3 or T4, Gleason score ≥8, and/or PSA ≥20 ng/ml at the time of diagnosis, based on criteria from the National Comprehensive Cancer Network (NCCN) Practice Guidelines (42,43). Low-risk PC was defined as clinical stage T1-T2 with Gleason score 2–6 and PSA <10 ng/ml, and intermediate-risk PC as T1-T2 with Gleason score 7 and/or PSA 10 to <20 ng/ml (42,43). All deaths attributed to PC as the primary cause were identified from the Swedish Cause of Death Registry using International Classification of Diseases (ICD) codes (ICD-7: 177, ICD-8/9: 185, ICD-10: C61). This registry includes deaths among all persons registered in Sweden since 1960, with compulsory reporting nationwide.

Covariates

Other variables that may be associated with CRF and PC were obtained from the Swedish Military Conscription Registry and national census data, which were linked using an anonymous personal identification number. The following were used as adjustment variables: birth date (modeled simultaneously as a continuous variable and categorical variable by decade); year of military conscription examination (continuous and categorical by decade); country of birth (Sweden/other); highest attained education level during the study period (≤9, 9–12, >12 years); first-degree family history of PC (yes or no, ascertained at age 50 years using NPCR and Swedish Cancer Registry data); and average neighborhood socioeconomic status (SES) during the study period (composed of an index that includes low education level, low income, unemployment, and social welfare receipt, as previously described (44), and categorized as low [>1 SD below the mean], medium [within 1 SD from the mean], or high [>1 SD above the mean]). Neighborhood SES was included because neighborhood deprivation has been associated with reduced CRF (45) as well as increased risk of PC diagnosis or mortality (46,47). Data were >99% complete for each variable. Missing data for each variable were coded as a separate category but had a negligible effect on the analyses because of their rarity.

Statistical Analysis

Poisson regression with robust standard errors was used to compute incidence rate ratios (IRRs) and 95% confidence intervals (CIs) for associations between CRF and subsequent risk of any PC, low- or intermediate-risk PC, aggressive PC, or PC mortality, examined in separate models. Two different adjustment models were used. The reduced model was adjusted only for birth date and year of the military conscription examination. The full model additionally included muscle strength, height, weight, country of birth, education level, neighborhood SES, and family history of PC. Poisson model goodness-of-fit was assessed using deviance and Pearson chi-squared tests, which showed a good fit in all models. Potential interactions between CRF and covariates were explored in relation to PC risk on both the multiplicative and additive scale. All statistical tests were 2-sided and used an α-level of 0.05. All analyses were conducted using Stata version 15.1.

In an exploratory subanalysis, we explored for evidence of detection bias (i.e., the possibility that men with high CRF are more likely to be diagnosed with PC because of increased screening) by assessing all ICD codes for cancer screening (ICD-9: V76, ICD-10: Z12; reported in 3,336 men) and specifically PC screening (ICD-10: Z12.5; 780 men) during the study period in the Swedish Outpatient and Primary Care Registries. The Outpatient Registry started in 2001 and contains outpatient diagnoses from all specialty clinics nationwide. The Primary Care Registry initially included all primary care diagnoses from two populous counties covering 20% of the national population starting in 1998, then was gradually expanded to cover >75% of the national population by 2008 and onward.

RESULTS

In 38.8 million person-years of follow-up, 10,782 (1.5%) men were diagnosed with PC, including 1,817 (0.3%) with aggressive PC and 217 (0.03%) who died from PC (mean follow-up time, 39.5 years). The median age at the end of follow-up was 57.0 years (mean 57.5 ± 4.2, range 50.0 to 75.0). The median age at any PC diagnosis was 56.8 years (mean 56.3 ± 4.3, range 37.2 to 66.7), at aggressive PC diagnosis was 57.4 years (mean 56.7 ± 4.4, range 39.9 to 65.6), and at PC death was 55.5 years (mean 54.7 ± 4.5, range 41.4 to 64.3). Participant characteristics are shown by CRF level in Table 1 and by subsequent PC diagnosis or mortality in Supplementary Table S1.

Table 1.

Characteristics of study participants by CRF quintile, 1972–1985, Sweden.

	CRF, Watts
	<219 (n=138,814) %	219–239 (n=130,843) %	240–260 (n=147,699) %	261–290 (n=141,382) %	≥291 (n=140,387) %
Age at baseline, years (mean ± SD)	18.5 ± 1.0	18.3 ± 0.9	18.3 ± 0.8	18.2 ± 0.7	18.2 ± 0.6
Muscle strength, Newtons (mean ± SD)	1893 ± 289	2014 ± 288	2085 ± 294	2148 ± 299	2220 ± 304
<1810	38.1	22.6	16.2	11.5	7.5
1810–1982	25.9	24.9	21.4	18.4	14.3
1983–2139	17.1	20.9	21.4	20.9	19.2
2140–2324	12.2	18.3	21.5	23.4	24.7
≥2325	6.7	13.2	19.5	5.8	34.2
Missing	<0.1	<0.1	<0.1	<0.1	0.1
Height, cm (mean ± SD)	176.5 ± 6.9	178.1 ± 6.6	179.0 ± 6.4	179.9 ± 6.4	181.3 ± 6.3
<174.0	32.0	23.3	19.1	14.9	10.1
174.0–176.9	18.0	17.3	15.9	14.6	12.1
177.0–180.9	23.1	24.9	25.0	24.8	23.4
181.0–183.9	12.7	15.1	16.5	17.8	19.0
≥184.0	14.2	19.4	23.5	27.9	35.4
Missing	<0.1	<0.1	<0.1	<0.1	<0.1
Weight, kg (mean ± SD)	63.4 ± 9.7	67.4 ± 9.8	69,7 ± 9.9	71.5 ± 9.7	73.7 ± 9.0
<61.0	41.9	21.8	13.4	7.9	3.7
61.0–65.9	25.4	26.5	22.4	18.4	12.3
66.0–69.9	13.5	18.8	20.1	20.0	17.6
70.0–75.9	10.4	17.6	22.7	26.8	30.5
≥76.0	8.8	15.3	21.4	26.9	35.8
Missing	<0.1	<0.1	<0.1	<0.1	0.1
BMI, kg/m2 (mean ± SD)	20.3 ± 2.8	21.2 ± 2.8	21.8 ± 2.8	22.1 ± 2.6	22.4 ± 2.4
Normal (<25.6)	95.4	93.3	91.7	91.2	91.5
Overweight (25.6–28.9)	3.0	4.6	5.9	6.5	6.6
Obesity (≥29.0)	1.6	2.1	2.4	2.2	1.8
Missing	<0.1	<0.1	<0.1	0.1	0.1
Born in Sweden
Yes	96.0	96.7	96.9	97.3	97.6
No	3.9	3.2	3.0	2.6	2.3
Missing	0.1	0.1	0.1	0.1	0.1
Education (years)
≤9	26.6	21.5	17.6	12.8	7.8
10–12	53.5	53.7	53.3	51.7	46.5
>12	19.9	24.8	29.0	35.5	45.7
Missing	<0.1	<0.1	<0.1	<0.1	<0.1
Neighborhood SES
Low	11.4	9.4	8.2	6.8	4.9
Medium	71.9	71.9	71.8	70.5	68.6
High	15.5	17.6	19.2	22.0	26.1
Missing	1.2	1.0	0.8	0.7	0.4
Family history of PC	15.0	15.0	14.8	14.7	14.8

BMI = body mass index, CRF = cardiorespiratory fitness, PC = prostate cancer, SES = socioeconomic status.

CRF Results

After adjusting for sociodemographic factors, height, weight, and family history, high CRF at age 18 years was associated with a slightly increased risk of any PC (e.g., highest vs. lowest quintile: IRR, 1.10; 95% CI, 1.03–1.19; P=0.008; Table 2). This increased risk was driven by an association with low- or intermediate-risk PC, which comprised 83% of PC cases (highest vs. lowest quintile: IRR, 1.12; 95% CI, 1.04–1.22; P=0.004; Table 2). However, CRF was not significantly associated with risk of aggressive PC (highest vs. lowest quintile: IRR, 1.01; 95% CI, 0.85–1.21; P=0.90) nor PC mortality (1.24; 0.73–2.13; P=0.42). The PC mortality risk estimates had low precision due to the small number of PC deaths (n=217). When low-risk and intermediate-risk PC were examined separately, similar associations with CRF were found (Supplementary Table S2).

Table 2.

Associations between CRF or other factors at age 18 years (1972–1985) and subsequent PC diagnosis or mortality (1998–2017), Sweden.

	Any PC (n=10,782)					Low- or intermediate-risk PC (n=8,965)a			Aggressive PC (n=1,817)b			PC mortality (n=217)
	Reduced Modelc		Full Modeld			Full Modeld			Full Modeld			Full Modeld
	IRR	95% CI	IRR	95% CI	P	IRR	95% CI	P	IRR	95% CI	P	IRR	95% CI	P
CRF (Watts, quintiles)
<219	1.00		1.00			1.00			1.00			1.00
219–239	1.06	1.00, 1.12	1.03	0.97, 1.09	0.31	1.05	0.98, 1.12	0.15	0.96	0.83, 1.09	0.51	0.85	0.55, 1.32	0.48
240–260	1.13	1.07, 1.19	1.07	1.01, 1.13	0.02	1.09	1.02, 1.17	0.007	0.97	0.84, 1.11	0.64	1.31	0.88, 1.93	0.18
261–290	1.15	1.08, 1.22	1.07	1.00, 1.14	0.04	1.08	1.01, 1.16	0.03	1.00	0.86, 1.16	0.99	1.28	0.82, 1.99	0.27
≥291	1.23	1.15, 1.31	1.10	1.03, 1.19	0.008	1.12	1.04, 1.22	0.004	1.01	0.85, 1.21	0.90	1.24	0.73, 2.13	0.42
Per 50 Watts (trend)	1.08	1.05, 1.11	1.03	1.01, 1.06	0.02	1.04	1.01, 1.07	0.008	1.00	0.93, 1.07	0.99	1.15	0.94, 1.40	0.19
Muscle strength (Newtons, quintiles)
<1810	1.00		1.00			1.00			1.00			1.00
1810–1982	1.06	1.00, 1.12	1.04	0.98, 1.11	0.17	1.06	1.00, 1.14	0.06	0.94	0.82, 1.09	0.41	0.87	0.57, 1.32	0.52
1983–2139	1.07	1.01, 1.14	1.05	0.99, 1.12	0.12	1.07	1.00, 1.15	0.05	0.95	0.82, 1.11	0.51	0.94	0.61, 1.46	0.79
2140–2324	1.12	1.06, 1.19	1.11	1.04, 1.18	0.002	1.13	1.05, 1.21	0.001	1.01	0.86, 1.17	0.95	0.90	0.57, 1.41	0.64
≥2325	1.14	1.07, 1.21	1.14	1.07, 1.23	<0.001	1.21	1.12, 1.30	<0.001	0.88	0.74, 1.04	0.14	0.81	0.48, 1.37	0.43
Per 300 Newtons (trend)	1.05	1.03, 1.07	1.06	1.04, 1.09	<0.001	1.08	1.06, 1.11	<0.001	0.97	0.92, 1.03	0.32	0.99	0.84, 1.18	0.92
Height (cm, quintiles)
<174.0	1.00		1.00			1.00			1.00			1.00
174.0–176.9	1.09	1.02, 1.16	1.05	0.99, 1.12	0.11	1.03	0.96, 1.11	0.42	1.18	1.01, 1.38	0.04	1.29	0.79, 2.10	0.30
177.0–180.9	1.12	1.05, 1.18	1.07	1.01, 1.13	0.03	1.07	1.00, 1.14	0.05	1.08	0.93, 1.25	0.33	1.62	1.05, 2.51	0.03
181.0–183.9	1.12	1.06, 1.20	1.07	1.00, 1.14	0.06	1.07	0.99, 1.15	0.07	1.05	0.89, 1.24	0.56	1.42	0.87, 2.33	0.16
≥184.0	1.15	1.09, 1.22	1.09	1.02, 1.16	0.009	1.06	0.98, 1.13	0.13	1.26	1.09, 1.49	0.003	1.75	1.10, 2.79	0.02
Per 5 cm (trend)	1.03	1.02, 1.05	1.02	1.01, 1.04	0.002	1.02	1.01, 1.04	0.006	1.03	0.99, 1.07	0.13	1.08	1.04, 1.12	<0.001
Weight (kg, quintiles)
<61.0	1.00		1.00			1.00			1.00			1.00
61.0–65.9	1.06	1.00, 1.12	0.98	0.92, 1.04	0.53	0.98	0.91, 1.05	0.49	1.00	0.86, 1.17	0.99	1.19	0.76, 1.87	0.44
66.0–69.9	1.12	1.05, 1.19	0.99	0.92, 1.06	0.71	0.97	0.90, 1.05	0.48	1.05	0.89, 1.24	0.54	1.06	0.65, 1.75	0.81
70.0–75.9	1.12	1.05, 1.19	0.96	0.89, 1.03	0.24	0.96	0.88, 1.03	0.25	0.97	0.81, 1.15	0.72	0.88	0.53, 1.46	0.62
≥76.0	1.02	0.95, 1.08	0.86	0.80, 0.93	<0.001	0.83	0.76, 0.91	<0.001	1.03	0.86, 1.25	0.72	0.95	0.55, 1.63	0.84
Per 5 kg (trend)	0.99	0.99, 1.00	0.96	0.95, 0.98	<0.001	0.95	0.94, 0.97	<0.001	1.01	0.99, 1.04	0.41	0.95	0.86, 1.05	0.34
BMI (kg/m2)e
Normal (<25.6)	1.00		1.00			1.00			1.00			1.00
Overweight (25.6–28.9)	0.86	0.78, 0.94	0.87	0.79, 0.95	0.003	0.83	0.75, 0.92	<0.001	1.07	0.86, 1.31	0.56	0.96	0.51, 1.78	0.89
Obesity (≥29.0)	0.62	0.52, 0.74	0.64	0.54, 0.77	<0.001	0.57	0.46, 0.70	<0.001	1.04	0.74, 1.47	0.82	0.49	0.12, 1.96	0.31
Per 5 BMI units (trend)	0.93	0.90, 0.96	0.89	0.85, 0.92	<0.001	0.86	0.83, 0.90	<0.001	1.01	0.92, 1.11	0.81	0.87	0.65, 1.17	0.37

^aLow-risk PC was defined as clinical stage T1-T2 with Gleason score 2–6 and PSA <10 ng/ml, and intermediate-risk PC as T1-T2 with Gleason score 7 and/or PSA 10 to <20 ng/ml.

^bAggressive PC was defined as clinical stage T3-T4, Gleason score ≥8, and/or PSA ≥20 ng/ml.

^cAdjusted for birth date and year of military conscription examination.

^dAdjusted for birth date, year of military conscription examination, CRF, muscle strength, height, weight, country of birth, education, neighborhood socioeconomic status, and family history of PC (except BMI was included as an alternative to height and weight in a separate model). The reference category for each variable is indicated by an IRR of 1.00.

^eDefined based on CDC criteria for males aged <20 years.

BMI = body mass index, CRF = cardiorespiratory fitness, IRR = incidence rate ratio, PC = prostate cancer.

Compared with IRRs in the reduced model, adjustment for covariates in the full model (as above) resulted in substantially lower IRRs for any PC (Table 2), but negligible change for aggressive PC and slightly higher IRRs for PC mortality (Supplementary Table S3 compared with Table 2). Education level was the strongest confounder that accounted for these changes. Figure 1 shows the probability of diagnosis with PC by CRF level, adjusted for covariates.

Figure 1.

Probability of diagnosis with PC (1998–2017) by cardiorespiratory fitness at age 18 years (1972–1985), adjusted for covariates.

Secondary Results

Similar to the results for CRF, high muscle strength was associated with a modestly increased risk of any PC (e.g., highest vs. lowest quintile: IRR, 1.14; 95% CI, 1.07–1.23; P<0.001), which was driven by its association with low- or intermediate-risk PC (1.21; 1.12–1.30; P<0.001). However, muscle strength was not associated with aggressive PC (highest vs. lowest quintile: IRR, 0.88; 95% CI, 0.74–1.04; P=0.14) nor PC mortality (0.81; 0.48–1.37; P=0.43).

Height was positively associated with risk of any PC, aggressive PC, and PC mortality. Compared with men in the lowest quintile for height, those in the highest quintile had a 9% increased risk of any PC (IRR, 1.09; 95% CI, 1.02–1.16; P=0.009), 26% increased risk of aggressive PC (1.26; 1.09–1.49; P=0.003), and 75% increased risk of PC mortality (1.75; 1.10–2.79; P=0.02) in the fully adjusted model.

In contrast, men in the highest quintile for weight at age 18 years had a slightly reduced subsequent risk of any PC compared with those in the lowest quintile (IRR, 0.86; 95% CI, 0.80–0.93; P<0.001), but no significant difference in risk of aggressive PC (1.03; 0.86–1.25; P=0.72) nor PC mortality (0.95; 0.55–1.63; P=0.84). Similarly, obesity at age 18 years was associated with a 36% lower risk of any PC (IRR, 0.64; 95% CI, 0.54–0.77; P<0.001), but was not significantly associated with risk of aggressive PC (1.04; 0.74–1.47; P=0.82) nor PC mortality (0.49; 0.12–1.96; P=0.31), compared with men who had a normal BMI at baseline.

Men with high education level or neighborhood SES had a significantly increased risk of low- or intermediate-risk PC, but not aggressive PC nor PC mortality. No interactions between CRF or muscle strength and any other variables were found in relation to PC risk or mortality. When stratifying by height (above vs. below the median), only modest differences in risk estimates were seen (Supplementary Table S4). No significant additive or multiplicative interactions were found between CRF and height (Supplementary Table S5) nor between muscle strength and height (Supplementary Table S6) in relation to risk of any PC, aggressive PC, or PC mortality.

Exploratory Subanalysis

The association between CRF and the reported prevalence of any cancer screening or PC screening was examined to explore for potential detection bias due to increased screening in men with high CRF. ICD codes for cancer screening or PC screening were reported for 3,336 and 780 men, respectively. Using these available data, men in the highest CRF quintile were significantly more likely to have cancer screening (IRR, 1.34; 95% CI, 1.20–1.49; P<0.001) or PC screening (1.37; 1.10–1.72; P=0.006), compared with those in the lowest quintile. These findings are consistent with the possibility of detection bias (i.e., the positive association observed between CRF and diagnosis with any PC may be due to increased screening among men with high CRF). Further adjustment for cancer screening or PC screening in the main analyses had a negligible effect on the risk estimates, possibly due to the scarcity of this information.

DISCUSSION

In this large population-based cohort, high CRF at age 18 years was associated with a modestly increased risk of any PC, and specifically low- or intermediate-risk disease. However, we found no significant association between CRF and risk of aggressive PC or PC mortality. The observed association with low- or intermediate-risk PC may be related to increased screening among men with high CRF, as suggested by our exploratory analysis of cancer screening. Also consistent with such an effect, we found that high education and neighborhood SES levels, which likely are associated with increased PC screening, were also associated with low- or intermediate-risk PC but not aggressive PC. These findings demonstrate the importance of distinguishing aggressive from indolent PC and assessing the possibility of detection bias in studies of PC risk factors.

Objectively measured CRF has rarely been examined in relation to overall PC risk (17–22), and to our knowledge never in relation to risk of aggressive PC. Prior studies of CRF in mid-adulthood in relation to overall PC have yielded conflicting results, possibly due to relatively small sample sizes and combining aggressive and indolent PC cases. A US study with 337 PC cases in 4,920 men (mean age 59 years) reported no significant association between CRF and overall PC risk, although there was a non-significant trend toward an inverse association (22). Another US study with 634 self-reported PC cases in 19,042 men (mean age 46 years) with an average follow-up of 9 years reported a positive association between CRF and overall PC risk only before 1995 (i.e., during the first few years after introduction of PSA screening in 1987) (18). This finding suggested the possibility of detection bias, wherein men with high CRF were more likely to be early adopters of PSA screening, leading to a spurious positive association with overall PC risk (18). A later study from the same cohort examined 1,310 PC cases at ages ≥65 years from Medicare records and reported similar findings, but also may have been susceptible to detection bias (19). A Finnish study with 127 PC cases in 2,268 men (mean age 53 years) (20) and a Norwegian study with 213 PC cases in 1,997 men (mean age 50 years) (21) reported no association between CRF and overall PC risk. All of these studies were based on sampling from selected clinics rather than national populations, and none examined the association between CRF and aggressive PC.

Other studies have suggested that self-reported high physical activity may be related to decreased risk of any PC. A meta-analysis that included 88,294 PC cases from 19 cohort studies and 24 case-control studies reported that high physical activity (variably defined and ascertained) was associated with a modestly reduced risk of any PC (pooled relative risk [RR] for highest vs. lowest category, 0.90; 95% CI, 0.84–0.95) (24). However, a review of meta-analyses also concluded that the existing epidemiologic data are inconsistent and do not provide strong evidence to support an association between physical activity and risk of PC (25).

Because aggressive PC is common and lethal (1,2), the identification of modifiable factors early in life is a public health priority. However, no modifiable behavioral risk factors have yet been identified, with the possible exception of BMI, for which evidence is inconsistent. Some (48,49) but not all (50–52) studies have suggested a modest association between high BMI and risk of advanced PC (variably defined). A meta-analysis of 12 studies reported conflicting findings for BMI in relation to localized vs. advanced PC risk: an inverse association with localized PC (RR per 5 unit increase in BMI, 0.94; 95% CI, 0.91–0.97) but a modest positive association with advanced PC (RR, 1.09; 1.02–1.16), with weak evidence for study heterogeneity (52). In contrast, we found that high BMI at age 18 years was associated with a reduced risk of any PC later in life, and was not significantly associated with aggressive PC or PC mortality. These conflicting findings need further elucidation in studies with longitudinal BMI measurements and the ability to distinguish aggressive from indolent PC.

To our knowledge, the present study is the first to examine objectively measured muscle strength in relation to future risk of PC. Similar to the results for CRF, muscle strength was associated with a modestly increased risk of any PC, but not aggressive PC nor PC mortality. Our findings also suggested that tall men have a significantly increased risk for these outcomes. Men in the highest compared to lowest quintile for height had an estimated 26% (95% CI, 9% to 49%) and 75% (10% to 179%) higher risk of aggressive PC or PC mortality, respectively. These findings are consistent with prior studies that have linked tall height with increased risk of PC (49,53) as well as other site-specific cancers (54). The underlying mechanisms are not established but may involve early exposure to growth factors such as IGF-I, which has been associated with PC risk (55).

Strengths of the present study include objective measurement of CRF and prospective ascertainment of PC and cancer characteristics in a large population-based cohort. This study design helped minimize potential selection bias. The use of registry data with prospectively measured CRF, height, weight, family history, and socioeconomic factors also avoided recall bias.

This study also had several limitations. CRF and muscle strength were measured only once at age 18 years, and hence we were unable to examine changes in these factors over time. Although these characteristics often persist into adulthood, longitudinal measurements are needed to further assess cumulative lifetime exposures. Misclassification of CRF or muscle strength through submaximal effort by military conscripts during the testing could have influenced results toward the null hypothesis. Information on timing of puberty or testosterone levels was unavailable. Earlier puberty has been hypothesized to increase the risk of PC due to increased length of time that the prostate is exposed to high levels of androgens (56), and may also be associated with higher CRF or muscle strength in adolescence by leading to greater lean muscle mass (57). Despite an average follow-up of nearly four decades, this was still a relatively young cohort. The mean age at end of follow-up was 57 years, more than 10 years younger than the median age at PC diagnosis in Sweden (58) or worldwide (2). These men also were substantially leaner at age 18 years compared with current young men, and thus generalizability to later cohorts is uncertain.

In summary, in this large population-based cohort, men with high CRF or muscle strength at age 18 years had a slightly increased risk of PC later in life, likely related to increased screening among men with these characteristics. However, neither CRF nor muscle strength was significantly associated with risk of aggressive PC or PC mortality. Future studies with information on other early-life environmental exposures are needed to explore other modifiable factors that potentially could enable men to reduce their lifetime PC risk. Such studies should distinguish aggressive from indolent PC and assess for potential detection bias.

Abbreviations:

BMI

body mass index

CDC

Centers for Disease Control and Prevention

CI

confidence interval

CRF

cardiorespiratory fitness

ICD

International Classification of Diseases

IGF

insulin-like growth factor

IRR

incidence rate ratio

NCCN

National Comprehensive Cancer Network

NPCR

National Prostate Cancer Register

PC

prostate cancer

PSA

prostate specific antigen

RR

relative risk

SD

standard deviation

SES

socioeconomic status

TNM

tumor, nodes, metastasis