Skip to main content
PMC home page
JNCI Cancer Spectrum logoLink to JNCI Cancer Spectrum
. 2024 Aug 24;8(5):pkae070. doi: 10.1093/jncics/pkae070

Adiposity assessed close to diagnosis and prostate cancer prognosis in the EPIC study

Margarita Cariolou 1, Sofia Christakoudi 2, Marc J Gunter 3, Tim Key 4, Aurora Pérez-Cornago 5, Ruth Travis 6, Raul Zamora-Ros 7, Kristina Elin T Petersen 8, Anne Tjønneland 9,10, Elisabete Weiderpass 11, Rudolf Kaaks 12, Petra Seibold 13, Elif Inan-Eroglu 14, Matthias B Schulze 15,16, Giovanna Masala 17, Claudia Agnoli 18, Rosario Tumino 19, Chiara Di Girolamo 20, Amaia Aizpurua 21,22, Miguel Rodriguez-Barranco 23,24,25, Carmen Santiuste 26,27, Marcela Guevara 28,29,30, Dagfinn Aune 31,32,33, Doris S M Chan 34, David C Muller 35, Konstantinos K Tsilidis 36,37,✉
PMCID: PMC11410200  PMID: 39180334

Abstract

Background

Adiposity has been characterized as a modifiable risk factor for prostate cancer. Its association with outcomes after prostate cancer diagnosis, however, must be better understood, and more evidence is needed to facilitate the development of lifestyle guidance for patients with prostate cancer.

Methods

We investigated the associations between adiposity indices close to prostate cancer diagnosis (up to 2 years before or up to 5 years after diagnosis) and mortality in 1968 men of the European Prospective Investigation into Cancer and Nutrition cohort. Men were followed up for a median of 9.5 years. Cox proportional hazards models were adjusted for age and year of diagnosis, disease stage and grade, and smoking history and stratified by country.

Results

Each 5-unit increment in prediagnosis or postdiagnosis body mass index combined was associated with a 30% higher rate of all-cause mortality and a 49% higher rate of prostate cancer–specific mortality. Similarly, each 5-unit increment in prediagnosis body mass index was associated with a 35% higher rate of all-cause mortality and a 51% higher rate of prostate cancer–specific mortality. The associations were less strong for postdiagnosis body mass index, with a lower number of men in analyses. Less clear positive associations were shown for waist circumference, hip circumference, and waist to hip ratio, but data were limited.

Conclusions

Elevated levels of adiposity close to prostate cancer diagnosis could lead to higher risk of mortality; therefore, men are encouraged to maintain a healthy weight. Additional research is needed to confirm whether excessive adiposity after prostate cancer diagnosis could worsen prognosis.

Prostate cancer is a major public health burden challenging the economic and health-care systems of many countries globally (1,2). It is the second-most frequently diagnosed malignancy after lung cancer and the fifth-leading cause of cancer-related deaths in men worldwide (1.7 million cases and 500 000 annual deaths expected by 2030) (3-5). Most men have a diagnosis of localized prostate cancer and up to 99% 10-year survival (6), whereas the 5-year survival for individuals with late-stage disease is approximately 30% (6). Survival rates have substantially improved (5,7,8), but extended survival often coexists with increased cancer-related comorbidities, including obesity (9).

General adiposity (body mass index [BMI]) has been inversely associated with the risk of developing localized and total prostate cancer (10-12); general adiposity (BMI) and abdominal adiposity (waist circumference, waist to hip ratio) have been positively associated with the risk of developing aggressive (10,11,13-16) and fatal prostate cancer (10,17) in studies, including European Prospective Investigation into Cancer and Nutrition (EPIC). The literature on potential associations between modifiable factors and prostate cancer prognosis is inconclusive (18-20). Currently, any existing lifestyle survivorship guidance is mainly extrapolated from cancer prevention recommendations. Individuals with cancer are advised to avoid obesity and adhere to a balanced lifestyle (20,21). A recent meta-analysis of observational studies reported evidence of a J-shaped association between BMI assessed at or after prostate cancer diagnosis and all-cause mortality (17 studies) and a similar association for prostate cancer–specific mortality (13 studies). No associations were identified between waist circumference and mortality (3 studies) (19). Observational studies that explored adiposity close to diagnosis in relation to prostate cancer survival are subject to limitations, such as being single centered, using a single-timepoint adiposity measure instead of measures at different timepoints throughout the cancer survivorship trajectory, focused on general adiposity without incorporating measures of abdominal or gluteofemoral (hip circumference) adiposity. A few studies investigated potential nonlinear relationships (22), but none explored stage-specific or grade-specific associations (19). Investigating the association between obesity and cancer-related survival outcomes should be among the top research priorities (23). We explored the associations among general, abdominal, and gluteofemoral adiposity assessed close to diagnosis with mortality in the EPIC study.

Methods

EPIC is a multicenter prospective cohort study that recruited volunteers aged 35-70 years between 1992 and 2000. The final cohort included 153 457 men. Study design and data-collection procedures are detailed elsewhere (24). Eligibility criteria, identification of prostate cancer cases, and outcome assessment are described in detail in the Supplementary Methods (available online). Briefly, men with prostate cancer were eligible for inclusion if they had adiposity data collected close to diagnosis (ie, up to 2 years before [prediagnosis] or up to 5 years after [postdiagnosis]) (Figure 1, A-D). Data were used from all EPIC centers apart from France, Norway, Utrecht, and Naples, which recruited only women. Data from Greece were not available for this analysis. The International Classification of Diseases for Oncology, 3rd Edition (code C61.9) was used to select individuals with primary malignant prostate tumors (25).

Figure 1.

Figure 1.

Open in a new tab

Data collection time frames showing the total number of men who had data on adiposity close to diagnosis for (A) BMI, (B) waist circumference, (C) hip circumference, and (D) waist to hip ratio. The prediagnosis analyses included men with incident prostate cancer diagnosed during cohort follow-up who had baseline (at recruitment) or follow-up anthropometric data collected up to 2 years before diagnosis. The postdiagnosis analyses included men with prevalent prostate cancer at study entry who had baseline anthropometric data collected up to 5 years after diagnosis or men with incident prostate cancer who had follow-up anthropometric data collected up to 5 years after diagnosis. The numbers shown on the diagram represent the total number of men with prostate cancer before exclusion of individuals with missing data on the covariates of the main model of the present study (ie, age at diagnosis, year of diagnosis, disease stage, tumor grade, and smoking status). BMI = body mass index.

Tumor grade and stage information at diagnosis was collected, if possible, from each center. Height, weight, waist, and hip circumference were obtained in all centers apart from Umea, which had information only about height and weight. Height, weight, hip, and waist circumference were measured using standard protocols in all participating EPIC centers in this study; only in Oxford, United Kingdom, were they self-reported. A follow-up assessment on average 5 years after recruitment collected data from 350 000 participants (28% men). Weight was self-reported in all centers apart from Norfolk, United Kingdom, where it was measured (24,26). At study entry, participants provided information about lifestyle, previous diseases, alcohol and tobacco consumption, exercise, diet, and sociodemographics. Supplementary Methods (available online) provides details of how important variables were defined.

Statistical analysis

Cox proportional hazards regression models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for the associations between adiposity assessed either up to 2 years before diagnosis or up to 5 years after diagnosis in relation to all-cause and prostate cancer–specific mortality. Analyses were also performed by combining both time frames. The date of the at-recruitment or follow-up anthropometric assessment/questionnaire was considered the start of follow-up, and the date of death, emigration, withdrawal or loss to follow-up or last follow-up, whichever occurred first, was the end of follow-up. In the analysis of prostate cancer–specific mortality; all other deaths were censored. We obtained estimates of the linear associations between adiposity variables and mortality per 5 kg/m2 increments in BMI, 10-cm increments in waist and hip circumference, and 0.1-unit increments in waist to hip-ratio. Restricted cubic splines were also used to investigate possible nonlinear associations, with 3 knots at the 10th, 50th, and 90th percentiles of the adiposity variable distribution (27). Nonlinearity was evaluated graphically and using the likelihood ratio test, comparing models with and without the restricted cubic spline term (28-30). The median value of each adiposity variable in each analysis was used as the referent.

Nested multivariable models were adjusted for relevant covariates selected a priori, on the basis of subject matter knowledge, including age and year of diagnosis, Gleason score/grade, tumor stage, prostate-specific antigen level, smoking status, and physical activity (Supplementary Table 1, available online). Model 1 was adjusted for age and year of diagnosis. Model 2 was additionally adjusted for disease stage and grade. Model 3, which represented our main model and is the one presented in our results, was additionally adjusted for smoking status. A fourth model, additionally adjusted for lifetime number of cigarettes per day, physical activity, and prostate-specific antigen level, resulted in similar results but a considerably larger percentage of missing data. All models were stratified by country to account for potential center-specific differences, including follow-up procedures and questionnaires (26,31). We performed complete-case analyses restricted to individuals with complete information about the covariates of each model (32,33). The number of men and deaths per model are shown in Supplementary Table 2 (available online). To examine the influence of covariates, we performed sensitivity analyses, including in models 1 and 2 the same participants as in model 3 (the main model). As a proxy for socioeconomic status, the prediagnosis or postdiagnosis BMI combined model was additionally adjusted for highest school level. The postdiagnosis models were also adjusted for prediagnosis BMI in sensitivity analysis. Analyses were conducted by the World Health Organization (WHO) BMI categories, but inference was the same (Supplementary Tables 3 and 4, available online).

Graphical inspection of the smoothed scaled Schoenfeld residuals showed no violation of the proportional hazards assumption (34-36). Sensitivity analyses were performed to assess potential reverse causation and selection bias (details in Supplementary Methods, available online). We also performed analyses by tumor stage and grade at diagnosis, smoking status, and year of BMI measurement (1 and 2 years before diagnosis and for each year after diagnosis) because each year may reflect differences in the cancer care continuum. A P value of .05 was considered statistically significant; statistical tests were 2 sided. R, version 4.0.5, software (R Foundation for Statistical Computing, Vienna, Austria) was used for analyses.

Results

Cohort characteristics

A total of 1968 men with prostate cancer had adiposity data collected up to 2 years before or up to 5 years after diagnosis. Important tumor and lifestyle characteristics of these men are shown in Table 1. Some data were missing for prostate-specific antigen (45%), tumor stage (39%), tumor grade (34%), lifetime cigarettes per day (31%), physical activity (15%), and smoking status (4%), but all men had complete information for age and year of diagnosis (Supplementary Table 5, available online). Tumor and baseline lifestyle characteristics of the men with prostate cancer who were not included in this study (if they did not have adiposity data close to diagnosis) were not materially different from those of the men who were included. Important characteristics of those included in the main BMI analyses (after removing men with missing covariate data) were comparable to the characteristics of the total number of men who had BMI data close to diagnosis (with missing data) (Supplementary Table 6, available online). Key characteristics of the 1968 men were also largely similar by the World Health Organization categories (Supplementary Table 7, available online).

Table 1.

Tumor and lifestyle characteristics of the men diagnosed with prostate cancer who had prediagnosis or postdiagnosis adiposity dataa

BMI (N = 1968) Waist circumference (n = 1091) Hip circumference (n = 793) Waist to hip ratio (n = 793)
Follow-up time (ie, from return of either baseline or follow-up questionnaire until censoring), median (5th-95th percentile), y 9.5 (2.0-17.6) 8.4 (2.0-18.3) 7.6 (2.0-18.8) 7.6 (2.0-18.8)
Age at diagnosis, median (5th-95th percentile), y 66 (55-77) 65 (54-74) 65 (54-75) 65 (54-75)
Disease stage—EPIC stage classification, No. (%)
 In situ 0 (0) 0 (0) 0 (0) 0 (0)
 Localized 710 (36) 308 (28) 116 (15) 116 (15)
 Metastatic 49 (3) 6 (1) 6 (1) 6 (1)
 Metastatic, regional 77 (4) 41 (4) 18 (2) 18 (2)
 Metastatic, distant 56 (3) 52 (5) 17 (2) 17 (2)
 Unknown/missing 1076 (55) 684 (63) 636 (80) 636 (80)
Tumor stage—TNM code or, if TNM not available, EPIC stage classification , No. (%)
 Localized (T0-T2 and N0-NX and M0) 824 (42) 314 (29) 123 (16) 123 (16)
 Advanced (T3-T4 and/or N1-N3 or M1) 382 (19) 191 (18) 129 (16) 129 (16)
 Unknown/missing 762 (39) 586 (54) 541 (68) 541 (68)
Tumor grade—EPIC grade classification, No. (%)
 Well differentiated 58 (3) 24 (2) 24 (3) 24 (3)
 Moderately differentiated 282 (14) 119 (11) 119 (15) 119 (15)
 Poor/undifferentiated 91 (5) 280 (26) 40 (5) 40 (5)
 Unknown/missing 1537 (78) 668 (61) 610 (77) 610 (77)
Tumor grade—Gleason score or, if not available, EPIC grade classification, No. (%)
 Gleason score 2-6 (well differentiated) 632 (32) 272 (25) 156 (20) 156 (20)
 Gleason score 7 (moderately differentiated) 456 (23) 208 (19) 146 (18) 146 (18)
 Gleason score 8-10 (poorly or undifferentiated) 211 (11) 117 (11) 65 (8) 65 (8)
 Unknown/undetermined 669 (34) 494 (45) 426 (54) 426 (54)
Tumor grade—Gleason score, otherwise EPIC grade classification (Gleason score 7 [moderately differentiated] as a separate category and split into 3 + 4 or 4 + 3), No. (%)
 Gleason score 2-6 (well differentiated) 632 (32) 272 (25) 156 (20) 156 (20)
 Gleason score 7 (3 + 4) (moderately differentiated) 199 (10) 91 (8) 50 (6) 50 (6)
 Gleason score 7 (4 + 3) (moderately differentiated) 78 (4) 22 (2) 18 (2) 18 (2)
 Gleason score 8-10 (poorly or undifferentiated) 211 (11) 117 (11) 65 (8) 65 (8)
 Unknown/undetermined 848 (43) 589 (54) 504 (64) 504 (64)
Anthropometry
BMI
 BMI, median (5th-95th percentile) 26 (22-33) 26 (22-33) 26 (22-33) 26 (22-33)
 Underweight (<18.5), No. (%) 8 (0) 5 (0) 5 (1) 5 (1)
 Normal weight (18.5-24.9), No. (%) 643 (33) 363 (33) 262 (33) 262 (33)
 Overweight (25-29.9), No. (%) 1037 (52) 572 (52) 414 (52) 414 (52)
 Obese (≥30), No. (%) 280 (14) 151 (14) 112 (14) 112 (14)
 Unknown, No. (%) —b —b —b —b
Waist circumference
 Waist circumference, median (5th-95th percentile) 97 (82-115) 97 (82-115) 97 (81-115) 97 (81-115)
 Unknown, No. (%) 877 (45) —b —b —b
Hip circumference
 Hip circumference, median (5th-95th percentile) 101 (90-114) 101 (90-114) 101 (90-114) 101 (90-114)
 Unknown, No. (%) 1175 (60) 298 (27) —b —b
Waist to hip ratio
 Waist to hip ratio, median (5th-95th percentile) 1 (0.85-1.06) 0.96 (0.85-1.06) 0.96 (0.85-1.06) 0.96 (0.85-1.06)
 Unknown, No. (%) 1175 (60) 298 (27) —b —b
Smoking status, No. (%)
 Never smoker 645 (33) 334 (31) 243 (31) 243 (31)
 Former smoker 979 (50) 537 (49) 397 (50) 397 (50)
 Current smoker 270 (14) 182 (17) 118 (15) 118 (15)
 Unknown 74 (4) 38 (3) 35 (4) 35 (4)
Cambridge Physical Activity Index, No. (%)
 Inactive 495 (25) 284 (26) 215 (27) 215 (27)
 Moderately inactive 591 (30) 396 (36) 298 (38) 298 (38)
 Moderately active 322 (16) 197 (18) 126 (16) 126 (16)
 Active 256 (13) 184 (17) 124 (16) 124 (16)
 Missing 304 (15) 30 (3) 30 (4) 30 (4)
Highest school level (baseline), No. (%)
 None 29 (2) 6 (1) 6 (1) 6 (1)
 Primary school completed 657 (33) 364 (33) 266 (34) 266 (34)
 Technical/professional school 433 (22) 249 (23) 174 (22) 174 (22)
 Secondary school 224 (11) 140 (13) 116 (15) 116 (15)
 Longer education (including university degree) 525 (27) 285 (26) 185 (23) 185 (23)
 Not specified 78 (4) 35 (3) 35 (4) 35 (4)
 Unknown 22 (1) 12 (1) 11 (1) 11 (1)
Prostate-specific antigen level, median (5th-95th percentile) 11 (3-134) 12 (3-249) 11 (2-166) 11.1 (2-166)
Open in a new tab
a

Percentages are rounded to the nearest whole number. The characteristics of individuals are taken from each respective time frame (either baseline or follow-up) for which the individual was selected according to the eligibility criteria of this study (ie, for any variables that also had follow-up measurements). BMI = body mass index; EPIC = European Prospective Investigation into Cancer and Nutrition.

b

Unknown or missing data.

Main results: BMI and mortality

The 1968 men with BMI data collected either up to 2 years before diagnosis (n = 1000) or up to 5 years after diagnosis (n = 968) (Figure 1, A; Supplementary Figure 1, A, available online) were followed for a median of 9.5 years from return of the baseline or follow-up questionnaire. Analysis of BMI assessed before or after diagnosis combined showed a linear increase in the rate of all-cause (HR per 5 kg/m2 = 1.30, 95% CI = 1.11 to 1.52, Pnonlinearity = .93) and prostate cancer–specific mortality (HR per 5 kg/m2 = 1.49, 95% CI = 1.21 to 1.84, Pnonlinearity = .11) (Table 2; Figure 2, A and B). A positive association was also observed between prediagnosis BMI and all-cause mortality (HR per 5 kg/m2 = 1.35, 95% CI = 1.11 to 1.65, Pnonlinearity = .71) as well as prostate cancer–specific mortality (HR per 5 kg/m2 = 1.51, 95% CI = 1.16 to 1.96, Pnonlinearity = .61) (Table 2; Figure 2, C and D). For postdiagnosis BMI and all-cause mortality, the 95% confidence intervals were wide, crossing the null (HR per 5 kg/m2 = 1.23, 95% CI = 0.91 to 1.66, Pnonlinearity = .48). Little evidence of nonlinearity was observed for prostate cancer–specific mortality because it was based on few events (Pnonlinearity = .04) (Table 2; Figure 2, E and F).

Table 2.

Cox proportional hazard ratios and 95% confidence intervals for the linear association between adiposity and all-cause and prostate cancer–specific mortality (complete case analyses—main model)

No. of events/total No. of men with prostate cancer Hazard ratio a (95% CI) P nonlinearity b
Main analyses
BMI (per 5 kg/m2)
 All-cause mortality
  Prediagnosis or postdiagnosis combined 320/942 1.30 (1.11 to 1.52) .93
  Prediagnosis 194/570 1.35 (1.11 to 1.65) .71
  Postdiagnosis 126/372 1.23 (0.91 to 1.66) .48
 Prostate cancer–specific mortality
  Prediagnosis or postdiagnosis combined 163/942 1.49 (1.21 to 1.84) .11
  Prediagnosis 100/570 1.51 (1.16 to 1.96) .61
  Postdiagnosis 63/372 1.74 (1.13 to 2.68) .04
Waist circumference (per 10 cm)
 All-cause mortality
  Prediagnosis or postdiagnosis combined 120/362 1.16 (0.97 to 1.39) .03
  Prediagnosis 86/245 1.31 (1.07 to 1.61) .06
  Postdiagnosis 34/117 0.67 (0.40 to 1.08) .72
 Prostate cancer–specific mortality
  Prediagnosis or postdiagnosis combined 79/362 1.26 (1.01 to 1.57) .03
  Prediagnosis 53/245 1.47 (1.14 to 1.89) .01
  Postdiagnosis 26/117 0.79 (0.46 to 1.37) .79
Hip circumference (per 10 cm)
 All-cause mortality
  Prediagnosis or postdiagnosis combined 45/167 1.12 (0.72 to 1.77) .88
  Prediagnosis 42/128 1.19 (0.74 to 1.91) .74
  Postdiagnosis 3/39 (Limited data) —c
 Prostate cancer–specific mortality
  Prediagnosis or postdiagnosis combined 23/167 1.48 (0.80 to 2.74) .29
  Prediagnosis 21/128 1.88 (0.94 to 3.77) .62
  Postdiagnosis 2/39 (Limited data) —c
Waist to hip ratio (per 0.1 unit)
 All-cause mortality
  Prediagnosis or postdiagnosis combined 45/167 1.18 (0.70 to 1.99) .47
  Prediagnosis 42/128 1.24 (0.71 to 2.15) .91
  Postdiagnosis 3/39 (Limited data) —c
 Prostate cancer–specific mortality
  Prediagnosis or postdiagnosis combined 23/167 1.47 (0.67 to 1.02) .50
  Prediagnosis 21/128 1.71 (0.67 to 4.34) .75
  Postdiagnosis 2/39 (Limited data) —c
Subgroup and sensitivity analyses
Prediagnosis or postdiagnosis BMI combined, per 5 kg/m2
 Disease staged
  All-cause mortality
   Localized 175/656 1.17 (0.92 to 1.49) .19
   Advanced (includes metastatic) 145/286 1.38 (1.09 to 1.73) .37
   All, excluding metastatic 233/819 1.25 (1.02 to 1.52) .26
   Only metastatic 87/123 1.37 (1.00 to 1.86) .19
  Prostate cancer–specific mortality
   Localized 61/656 1.45 (1.01 to 2.08) .13
   Advanced (includes metastatic) 102/286 1.52 (1.15 to 2.02) .39
   All, excluding metastatic 85/819 1.57 (1.16 to 2.13) .23
   Only metastatic 78/123 1.35 (0.97 to 1.87) .43
 Tumor gradee
  All-cause mortality
   Well differentiated—Gleason score 2-6 106/427 1.34 (1.02 to 1.72) .85
   Moderately differentiated—Gleason score 7 118/359 1.20 (0.89 to 1.62) .12
   Poorly differentiated—Gleason score 8-10 96/156 1.21 (0.88 to 1.65) .22
  Prostate cancer–specific mortality
   Well differentiated—Gleason score 2-6 34/427 1.49 (0.99 to 2.27) .12
   Moderately differentiated—Gleason score 7 59/359 1.33 (0.89 to 1.99) .71
   Poorly differentiated—Gleason score 8-10 70/156 1.51 (1.06 to 2.16) .19
Additional adjustment for baseline highest school level (proxy for socioeconomic status)
All-cause mortality 318/936 1.25 (1.06 to 1.47) .65
Prostate cancer–specific mortality 163/936 1.43 (1.15 to 1.77) .07
Lagged analysis—removing deaths during the first year of follow-up
All-cause mortality 308/930 1.34 (1.14 to 1.58) .90
Prostate cancer–specific mortality 155/930 1.54 (1.25 to 1.91) .10
Postdiagnosis BMI analyses, per 5 kg/m2
 Lagged analysis—removing deaths during the first year of follow-up
  All-cause mortality 117/363 1.34 (0.98 to 1.83) .44
  Prostate cancer–specific mortality 57/363 1.96 (1.24 to 3.08) .04
Additional adjustment for prediagnosis/baseline BMI
All-cause mortality 126/372 1.04 (0.55 to 1.99) .46
Prostate cancer–specific mortality 63/372 1.49 (0.61 to 3.67) .03
By each year of BMI measurement (up to 2 y before diagnosis and up to 5 y after diagnosis), per 5 kg/m2
First year before diagnosis
 All-cause mortality 78/239 1.07 (0.76 to 1.50) .52
 Prostate cancer–specific mortality 37/239 1.51 (0.91 to 2.50) .79
Second year before diagnosis
 All-cause mortality 116/331 1.59 (1.24 to 2.04) .80
 Prostate cancer–specific mortality 63/331 1.54 (1.11 to 2.15) .71
First year after diagnosis
 All-cause mortality 52/146 1.44 (0.84 to 2.46) .68
 Prostate cancer–specific mortality 30/146 1.84 (0.90 to 3.74) .03
Second year after diagnosis
 All-cause mortality 40/123 1.58 (0.90 to 2.76) .42
 Prostate cancer–specific mortality 20/123 2.09 (0.95 to 4.58) .93
Third year after diagnosis
 All-cause mortality 26/77 1.05 (0.52 to 2.13) .67
 Prostate cancer–specific mortality 10/77 1.91 (0.49 to 7.47) .91
Open in a new tab
a

Main model (model 3) hazard ratio adjusted for age, year of diagnosis, disease stage, tumor grade, and smoking status and stratified by EPIC country. BMI = body mass index; CI = confidence interval; EPIC = European Prospective Investigation into Cancer and Nutrition.

b

Analysis of variance test of models with the restricted cubic spline term compared with models with the linear term (without the spline term).

c

No available data.

d

Not adjusted for disease stage apart from the analysis that excludes men with metastatic prostate cancer.

e

Not adjusted for tumor grade.

Figure 2.

Figure 2.

Open in a new tab

Hazard ratios from Cox proportional hazards models with restricted cubic spline curves describing the association between BMI data collected before or after diagnosis combined and (A) all-cause mortality (men/deaths = 972/320); (B) prostate cancer–specific mortality (men/deaths = 972/163), prediagnosis BMI; (C) all-cause mortality (men/deaths = 570/194); (D) prostate cancer–specific mortality (men/deaths = 570/100), postdiagnosis BMI; (E) all-cause mortality (men/deaths = 372/126); (F) prostate cancer–specific mortality (men/deaths = 372/63). Hazard ratios are based on the main model, adjusted for age at diagnosis, year of diagnosis, disease stage, tumor grade, and smoking status; knots at the 10th, 50th, and 90th percentiles of BMI. The median BMI of the individuals included in analyses was used as the referent: 26.2 in the prediagnosis or postdiagnosis BMI analysis, 26.5 in the prediagnosis BMI analysis, and 25.9 in the postdiagnosis BMI analysis. The smooth density plot represents the density of the population across the spline variable. BMI = body mass index; CI = confidence interval.

Subgroup and sensitivity analyses for BMI and mortality

Of 1968 men, 42% (n = 824) had localized disease, 19% (n = 382) had advanced disease, 9% (n = 180) had metastatic disease, 39% (n = 762) had missing data, 32% (n = 632) had well-differentiated disease (Gleason scores 2-6), 23% (n = 456) had moderately differentiated disease (Gleason score 7), 11% (n = 211) had poorly or undifferentiated disease (Gleason scores 8-10), and 34% (n = 669) had missing data (Table 1). Positive and similar associations were observed between 5 kg/m2 increments of BMI and all-cause and prostate cancer–specific mortality for localized, advanced, and metastatic prostate cancer, after excluding metastatic tumors (Table 2; Supplementary Figure 2, available online), and across tumor grades (Table 2; Supplementary Figure 3, available online). Similar associations to the main analyses were observed upon exclusion of men who had died during the first year of follow-up (Table 2; Supplementary Figure 4, available online). In addition, adjusting the postdiagnosis model for baseline or prediagnosis BMI did not materially change the results (Supplementary Figure 5, available online).

Supplementary Figure 6 (available online) shows the number of men with BMI data at 1 and 2 years before diagnosis and at each year after diagnosis up to 5 years later. During the first year before diagnosis, the association between BMI and all-cause mortality appeared linear (HR per 5 kg/m2 = 1.07, 95% CI = 0.76. to 1.50, Pnonlinearity = .52) and similar for prostate cancer–specific mortality, but the 95% confidence intervals crossed the null (HR per 5 kg/m2 = 1.51, 95% CI = 0.91 to 2.50, Pnonlinearity = .79) (Table 2; Supplementary Figure 7, A and B, available online). A stronger positive association was seen for BMI assessed in the second year before diagnosis with all-cause mortality (HR per 5 kg/m2 = 1.59, 95% CI = 1.24 to 2.04, Pnonlinearity = .80) and prostate cancer–specific mortality (HR per 5 kg/m2 = 1.54, 95% CI = 1.11 to 2.15, Pnonlinearity = .71) (Table 2; Supplementary Figure 7, C and D, available online). There were indications for linearity for the first 3 years after diagnosis (Supplementary Figure 8, available online), but the 95% confidence intervals were wide, crossing the null. Data were scarce beyond the third year after diagnosis (plots not shown).

Stratified analysis by smoking status showed positive associations among never smokers (HR per 5 kg/m2 = 1.62, 95% CI = 1.24 to 2.14) and current smokers (HR per 5 kg/m2 = 1.45, 95% CI = 1.07 to 1.98) and no association among former smokers (HR per 5 kg/m2 = 0.98, 95% CI = 0.76 to 1.28) for all-cause mortality (Supplementary Table 8, Supplementary Figure 9, available online). Additional adjustment for education gave similar positive associations as the main analysis (Table 2; Supplementary Figure 10, available online). Results were similar with individuals in model 3, included in models 1 and 2 (Supplementary Table 9, available online).

Waist circumference, hip circumference, waist to hip ratio, and mortality

Waist circumference data collected up to 2 years before diagnosis (n = 547) or up to 5 years after diagnosis (n = 544) were available for 1091 of the 1968 men (Figure 1, B; Supplementary Figure 1, B, available online) followed for a median of 8.4 years. Analysis of prediagnosis or postdiagnosis waist circumference combined showed a linear increase in the rate of all-cause mortality (Pnonlinearity = .03) and prostate cancer–specific mortality (Pnonlinearity = .03) up to approximately 100 cm, with limited data after this point. Results were similar for prediagnosis waist circumference but less clear in the postdiagnosis analysis because of limited data (Table 2; Supplementary Figure 11, available online). Of 1968 men, 793 had hip circumference and waist to hip ratio data up to 2 years before diagnosis (n = 390) or up to 5 years after diagnosis (n = 403) (Figure 1, C and D; Supplementary Figure 1, C and D, available online), followed for a median of 7.6 years. There were indications of positive associations in the prediagnosis and postdiagnosis groups combined and the prediagnosis analyses for all-cause and prostate cancer–specific mortality, but the 95% confidence intervals were wide, crossing the null (Table 2; Supplementary Figures 12 and 13, available online). Data were scarce and not shown for postdiagnosis hip circumference and waist to hip ratio.

Discussion

In this study, each 5-unit increment in prediagnosis or postdiagnosis BMI combined was associated with a 30% higher rate of all-cause mortality and a 49% higher rate of prostate cancer–specific mortality, independent of tumor grade, disease stage, smoking status, and other confounders. Each 5-unit increment in prediagnosis BMI was associated with a 35% higher rate of all-cause mortality and a 51% higher rate of prostate cancer–specific mortality. The associations were less strong for postdiagnosis BMI because of a lower number of men in analyses. Data on waist circumference, hip circumference, and waist to hip ratio were more limited than for BMI, but there were indications for positive associations with mortality.

Observational studies in men with prostate cancer that investigated the associations between postdiagnosis adiposity and mortality found inconsistent results. Some (18,22,37) reported positive associations and others reported inverse associations (38-41) between BMI and mortality outcomes. Most linear dose-response meta-analyses in patients with prostate cancer reported no associations (95% CIs crossing the null) between at diagnosis/postdiagnosis BMI (42-44) and prostate cancer–specific or all-cause mortality; only 1 study reported a small increase in the rate of all-cause mortality (44). Our meta-analysis of 2023 identified a J-shaped association between postdiagnosis BMI and all-cause and prostate cancer–specific mortality. Most studies adjusted for stage, tumor grade, and treatment but not for smoking status (19). In this study, we observed less clear associations for BMI assessed up to 5 years after diagnosis with mortality (only 372 men in this analysis). Little evidence of nonlinearity was seen for prostate cancer–specific mortality on the basis of few events, but we observed positive associations between BMI assessed up to 2 years before diagnosis and mortality. We hypothesized that adiposity in the 2 years before diagnosis would be representative of the adiposity level at diagnosis because many men are diagnosed with localized prostate cancer that is usually asymptomatic (5,6). Subgroup analyses, such as by disease stage, tumor grade, or smoking status, by year of BMI measurement generally showed consistent positive associations. We found positive associations between BMI and mortality that were stronger in never smokers than in current and former smokers, consistent with studies that observed stronger associations for BMI and risk of advanced prostate cancer among never smokers (45-48).

Few studies to date have investigated the relationship between at diagnosis/postdiagnosis waist circumference (49-51) in patients with prostate cancer, but found no associations with mortality. In addition, thus far, no studies in patients with prostate cancer have investigated waist to hip ratio or hip circumference after diagnosis in relation to long-term survival outcomes. Our results indicated positive associations between waist circumference, hip circumference, waist to hip ratio, and mortality, although data were limited.

The biological mechanisms linking obesity to poor survival outcomes after prostate cancer diagnosis have not been fully elucidated, and additional research is required (52). Numerous metabolic imbalances and interrelated pathways, including altered sex steroid hormones, serum insulin levels, and free insulin-like growth factor 1 levels may influence prostate cancer prognosis (52-54). Obesity-induced hormone and inflammatory changes could facilitate tumor progression (37,55) as well as higher risk of metastasis (56-58) and death (37,59). Chronically elevated insulin levels could facilitate tumor progression (5,60) and development of comorbidities, including cardiovascular disease (61), a major cause of death in patients with prostate cancer (55). Obesity has been positively associated with higher neutrophil to lymphocyte ratio (62,63), and strong or highly suggestive evidence supports that high neutrophil to lymphocyte ratio is associated with worse survival outcomes in patients with prostate cancer (64). An elevated obesity-induced inflammatory environment could worsen treatment-related side effects, for example, leading to more severe or extended post-treatment cancer-related fatigue (65). Chronically elevated levels of glucose in blood can activate the insulin pathway and facilitate tumor progression, the repair of tumor cells after radiation therapy, or treatment resistance (65,66). Androgen-deprivation therapy is commonly given in fixed doses, irrespective of the body’s surface area, and this practice could result in insufficient testosterone suppression in men with obesity compared with men who have normal weight (67). Inadequate pharmacological castration has been associated with more aggressive tumor biology (68), higher risk of progression and metastasis, and higher rates of mortality after androgen-deprivation therapy (67,69). The influence of castration therapies at the molecular level needs to be better understood. Obesity-related gene transcription in the adipose tissue surrounding the prostate could trigger inflammatory and metabolic changes (eg, altered hormone homeostasis, altered tissue lipid composition) that could partially counteract the beneficial effects of castration therapies (70).

The strengths of this study include its prospective design, representation of men from 7 countries, detailed information about adiposity indices close to diagnosis, mortality outcomes, and confounders. Important sensitivity analyses were performed, such as by distinct subgroups of disease stage and tumor grade. The period after cancer diagnosis, during and after treatment, is particularly complex because it involves various biological, behavioral, and physiological changes (71). Involuntary lifestyle alterations related to the disease and its treatment could influence body composition and in turn negatively affect cancer outcomes (9,71-74). We did not have information about cancer treatment, complications, or disease recurrence. Adjusting for diagnosis year, disease stage, and tumor grade could have potentially mitigated the limitation of not having treatment information because these variables reflect, to some extent, the treatment received and improvements in available prostate cancer therapies over time (75-78). Lack of repeated postdiagnosis adiposity measurements did not enable us to examine how cumulative changes in adiposity (and other lifestyle factors) could influence mortality by performing time-varying analyses. Our analysis of prediagnosis or postdiagnosis adiposity combined, however, has (at least partly) accounted for this limitation. Some men did not have data on the confounders of our main model (mainly, disease stage and tumor grade), particularly for the postdiagnosis time frame. All methods to account for missing data have limitations, and we acknowledge that performing a complete-case analysis was rather simplistic but likely the best possible approach to deal with the missing data in this study. We have provided detailed information about data missingness for transparency. A missing-indicator analysis could lead to biased estimates (79), and multiple imputation could have been more useful, with smaller amounts of missing data across variables (32). We investigated potential selection bias, but no material differences were observed in important lifestyle and tumor characteristics of the men included in and excluded from the analyses because of missing data.

Conclusions and future directions

Maintaining a healthy weight could lead to better prostate cancer prognosis. Additional well-designed and well-conducted observational and weight management intervention studies with larger sample sizes and repeated postdiagnosis measurements are needed. Mechanistic studies are essential to unravel the biological pathways driving tumor progression and mortality according to obesity-related factors and ultimately enable the design of targeted lifestyle interventions that will help men better cope with this disease.

Supplementary Material

pkae070_Supplementary_Data
pkae070_supplementary_data.pdf (2.4MB, pdf)

Acknowledgements

The funder did not play a role in the design of the study; the collection, analysis, and interpretation of the data; the writing of the manuscript; or the decision to submit the manuscript for publication.

We thank all participants in the EPIC cohort for their contribution to the study. We acknowledge the use of data from the EPIC-Bilthoven cohort (principle investigator W. M. Monique Verschuren). EPIC-Bilthoven thanks the National Institute for Public Health and the Environment, Bilthoven, the Netherlands, for their contribution and ongoing support to the EPIC Study. We acknowledge the use of data from the EPIC-Norfolk cohort (principle investigator Nick Wareham). EPIC-Norfolk thanks all the participants who have been part of the project and the many members of the study teams at the University of Cambridge who have enabled this research. We also acknowledge the use of data from the EPIC-Asturias cohort (principle investigator José Ramón Quirós), EPIC-Aarhus cohort (principle investigator Christina C. Dahm), EPIC-Malmo cohort (principle investigator Jonas Manjer), and EPIC-Umea cohort (principle investigator Bethany van Guelpen).

Disclaimer: Where authors are identified as personnel of the International Agency for Research on Cancer/WHO, the authors alone are responsible for the views expressed in this article, and they do not necessarily represent the decisions, policy, or views of the International Agency for Research on Cancer/WHO.

Contributor Information

Margarita Cariolou, Department of Epidemiology and Biostatistics, School of Public Health, Faculty of Medicine, Imperial College London, London, UK.

Sofia Christakoudi, Department of Epidemiology and Biostatistics, School of Public Health, Faculty of Medicine, Imperial College London, London, UK.

Marc J Gunter, Department of Epidemiology and Biostatistics, School of Public Health, Faculty of Medicine, Imperial College London, London, UK.

Tim Key, Cancer Epidemiology Unit, Nuffield Department of Population Health, University of Oxford, Oxford, UK.

Aurora Pérez-Cornago, Cancer Epidemiology Unit, Nuffield Department of Population Health, University of Oxford, Oxford, UK.

Ruth Travis, Cancer Epidemiology Unit, Nuffield Department of Population Health, University of Oxford, Oxford, UK.

Raul Zamora-Ros, Unit of Nutrition and Cancer, Cancer Epidemiology and Research Programme, Catalan Institute of Oncology (ICO), Bellvitge Biomedical Research Institute (IDIBELL), Barcelona, Spain.

Kristina Elin T Petersen, Danish Cancer Institute, Diet, Cancer and Health, Copenhagen, Denmark.

Anne Tjønneland, Danish Cancer Institute, Diet, Cancer and Health, Copenhagen, Denmark; Department of Public Health, University of Copenhagen, Copenhagen , Denmark.

Elisabete Weiderpass, International Agency for Research on Cancer, Lyon, France.

Rudolf Kaaks, Division of Cancer Epidemiology, German Cancer Research Center, Heidelberg, Germany.

Petra Seibold, Division of Cancer Epidemiology, German Cancer Research Center, Heidelberg, Germany.

Elif Inan-Eroglu, Department of Molecular Epidemiology, German Institute of Human Nutrition Potsdam-Rehbruecke, Nuthetal, Germany.

Matthias B Schulze, Department of Molecular Epidemiology, German Institute of Human Nutrition Potsdam-Rehbruecke, Nuthetal, Germany; Institute of Nutritional Science, University of Potsdam, Nuthetal, Germany.

Giovanna Masala, Clinical Epidemiology Unit, Institute for Cancer Research, Prevention, and Clinical Network, Florence, Italy.

Claudia Agnoli, Epidemiology and Prevention Unit, IRCCS National Cancer Institute Foundation, Milan, Italy.

Rosario Tumino, Hyblean Association for Epidemiological Research, AIRE ONLUS Ragusa, Ragusa, Italy.

Chiara Di Girolamo, Centre for Biostatistics, Epidemiology, and Public Health, Department of Clinical and Biological Sciences, University of Turin, Orbassano, Italy.

Amaia Aizpurua, Ministry of Health of the Basque Government, Sub directorate for Public Health and Addictions of Gipuzkoa, San Sebastián, Spain; Biodonostia Health Research Institute, Epidemiology of Chronic and Communicable Diseases Group, San Sebastián, Spain.

Miguel Rodriguez-Barranco, Escuela Andaluza de Salud Pública, Granada, Spain; Instituto de Investigación Biosanitaria ibs. GRANADA, Granada, Spain; Centro de Investigación Biomédica en Red de Epidemiología y Salud Pública, Madrid, Spain.

Carmen Santiuste, Centro de Investigación Biomédica en Red de Epidemiología y Salud Pública, Madrid, Spain; Department of Epidemiology, Murcia Regional Health Council, Murcia-IMIB, Spain.

Marcela Guevara, Centro de Investigación Biomédica en Red de Epidemiología y Salud Pública, Madrid, Spain; Instituto de Salud Pública y Laboral de Navarra, Pamplona, Spain; Navarra Institute for Health Research, Pamplona, Spain.

Dagfinn Aune, Department of Epidemiology and Biostatistics, School of Public Health, Faculty of Medicine, Imperial College London, London, UK; Department of Nutrition, Oslo New University College, Oslo, Norway; Department of Research, Cancer Registry of Norway, Norwegian Institute of Public Health, Oslo, Norway.

Doris S M Chan, Department of Epidemiology and Biostatistics, School of Public Health, Faculty of Medicine, Imperial College London, London, UK.

David C Muller, Department of Epidemiology and Biostatistics, School of Public Health, Faculty of Medicine, Imperial College London, London, UK.

Konstantinos K Tsilidis, Department of Epidemiology and Biostatistics, School of Public Health, Faculty of Medicine, Imperial College London, London, UK; Department of Hygiene and Epidemiology, University of Ioannina Medical School, Ioannina, Greece.

Data availability

For information about how to apply for gaining access to EPIC data or biospecimens, please follow the instructions at https://epic.iarc.fr/access.

Author contributions

Margarita Cariolou, MPH (Conceptualization; Data curation; Formal analysis; Writing – original draft); Doris S. M. Chan, PhD (Conceptualization; Funding acquisition; Resources; Supervision; Writing – review & editing); Dagfinn Aune, PhD (Writing – review & editing); Marcela Guevara, MD, PhD (Writing – review & editing); Carmen Santiuste, MD (Writing – review & editing); Miguel Rodriguez-Barranco, PhD (Writing – review & editing); Amaia Aizpurua, MSc (Writing – review & editing); Chiara Di Girolamo, PhD (Writing – review & editing); Rosario Tumino, MD (Writing – review & editing); Claudia Agnoli, MSc (Writing – review & editing); Giovanna Masala, MD (Writing – review & editing); Matthias B. Schulze, DrPH (Writing – review & editing); Elif Inan-Eroglu, PhD (Writing – review & editing); Petra Seibold, PhD (Writing – review & editing); Rudolf Kaaks, PhD (Writing – review & editing); Elisabete Weiderpass, MD, PhD (Writing – review & editing); Anne Tjønneland, MD, PhD (Writing – review & editing); Kristina Elin T. Petersen, PhD (Writing – review & editing); Raul Zamora-Ros, PhD (Writing – review & editing); Ruth Travis, DPhil (Writing – review & editing); Aurora Pérez-Cornago, PhD (Writing – review & editing); Tim Key, DPhil (Writing – review & editing); Marc J. Gunter, PhD (Writing – review & editing); Sofia Christakoudi, PhD (Methodology; Resources; Writing – review & editing); David C. Muller, PhD (Methodology; Supervision; Visualization; Writing – review & editing); Konstantinos K. Tsilidis, PhD (Conceptualization; Funding acquisition; Methodology; Project administration; Resources; Supervision; Visualization; Writing – review & editing).

Funding

This work was supported by the World Cancer Research Fund network of charities (American Institute for Cancer Research, World Cancer Research Fund, and Wereld Kanker Onderzoek Fonds). The coordination of EPIC is financially supported the International Agency for Research on Cancer and by the Department of Epidemiology and Biostatistics, School of Public Health, Imperial College London, which has additional infrastructure support provided by the National Institute for Health and Care Research Imperial Biomedical Research Centre. The national cohorts are supported by the Danish Cancer Society; German Cancer Aid, German Cancer Research Center, German Institute of Human Nutrition Potsdam-Rehbruecke, Federal Ministry of Education and Research (Germany); Associazione Italiana per la Ricerca sul Cancro-AIRC-Italy, Compagnia di SanPaolo, and National Research Council (Italy); Health Research Fund, Institute of Health Carlos III, and the regional governments of Andalucía, Asturias, Basque Country, Murcia, and Navarra, and the Catalan Institute of Oncology (Spain); and Cancer Research UK (C864/A14136 to EPIC-Norfolk; C8221/A29017 to EPIC-Oxford), Medical Research Council (MR/N003284/1, MC-UU_12015/1 and MC_UU_00006/1 to EPIC-Norfolk; MR/M012190/1 to EPIC-Oxford) (United Kingdom).

R. Zamora-Ros was supported by the Miguel Servet II (CPII20/00009) program from Institute of Health Carlos III (co-funded by the European Social Fund investing in your future).

The open access fee was paid from the Imperial College London Open Access Fund.

Conflicts of interest

The authors have no conflicts of interest.

References

  • 1. Ellinger J, Alajati A, Kubatka P, et al. Prostate cancer treatment costs increase more rapidly than for any other cancer-how to reverse the trend? EPMA J. 2022;13(1):1-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2. Pernar CH, Ebot EM, Wilson KM, Mucci LA.. The epidemiology of prostate cancer. Cold Spring Harb Perspect Med. 2018;8(12):a030361. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3. Sung H, Ferlay J, Siegel RL, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2021;71(3):209-249. [DOI] [PubMed] [Google Scholar]
  • 4. Bashir MN. Epidemiology of prostate cancer. Asian Pac J Cancer Prev. 2015;16(13):5137-5141. [DOI] [PubMed] [Google Scholar]
  • 5. Rawla P. Epidemiology of prostate cancer. World J Oncol. 2019;10(2):63-89. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6. Rebello RJ, Oing C, Knudsen KE, et al. Prostate cancer. Nat Rev Dis Primers. 2021;7(1):9. [DOI] [PubMed] [Google Scholar]
  • 7. Barsouk A, Padala SA, Vakiti A, et al. Epidemiology, staging and management of prostate cancer. Med Sci (Basel). 2020;8(3):28. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Cancer Research UK. https://www.cancerresearchuk.org/health-professional/cancer-statistics/statistics-by-cancer-type/prostate-cancer/incidence#heading-Three. Accessed December 20, 2020.
  • 9. Anderson AS, Martin RM, Renehan AG, et al. ; UK NIHR Cancer and Nutrition Collaboration (Population Health Stream). Cancer survivorship, excess body fatness and weight-loss intervention-where are we in 2020? Br J Cancer. 2020;124(6):1057-1065. doi: 10.1038/s41416-020-01155-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10. Perez-Cornago A, Appleby PN, Pischon T, et al. Tall height and obesity are associated with an increased risk of aggressive prostate cancer: results from the EPIC cohort study. BMC Med. 2017;15(1):115. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11. Discacciati A, Orsini N, Wolk A.. Body mass index and incidence of localized and advanced prostate cancer—a dose-response meta-analysis of prospective studies. Ann Oncol. 2012;23(7):1665-1671. [DOI] [PubMed] [Google Scholar]
  • 12. Kazmi N, Haycock P, Tsilidis K, et al. ; PRACTICAL Consortium, CRUK, BPC3, CAPS, PEGASUS. Appraising causal relationships of dietary, nutritional and physical-activity exposures with overall and aggressive prostate cancer: Two-sample Mendelian-randomization study based on 79 148 prostate-cancer cases and 61 106 controls. Int J Epidemiol. 2020;49(2):587-596. [DOI] [PubMed] [Google Scholar]
  • 13. Pischon T, Boeing H, Weikert S, et al. Body size and risk of prostate cancer in the European prospective investigation into cancer and nutrition. Cancer Epidemiol Biomarkers Prev. 2008;17(11):3252-3261. [DOI] [PubMed] [Google Scholar]
  • 14. Guerrios-Rivera L, Howard L, Frank J, et al. Is body mass index the best adiposity measure for prostate cancer risk? Results from a veterans affairs biopsy cohort. Urology. 2017;105:129-135. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15. Cao Y, Giovannucci E.. Obesity and prostate cancer. Recent Results Cancer Res. 2016;208:137-153. [DOI] [PubMed] [Google Scholar]
  • 16. World Cancer Research Fund International/American Institute for Cancer Research Continuous Update Project Report. Diet, Nutrition, Physical Activity, and Prostate Cancer. 2018. https://www.wcrf.org/sites/default/files/Prostate-cancer-report.pdf. Accessed December 10, 2023.
  • 17. Perez-Cornago A, Dunneram Y, Watts EL, Key TJ, Travis RC.. Adiposity and risk of prostate cancer death: a prospective analysis in UK Biobank and meta-analysis of published studies. BMC Med. 2022;20(1):143. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18. Troeschel AN, Hartman TJ, Jacobs EJ, et al. Postdiagnosis body mass index, weight change, and mortality from prostate cancer, cardiovascular disease, and all causes among survivors of nonmetastatic prostate cancer. J Clin Oncol. 2020;38(18):2018-2027. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19. Cariolou M, Markozannes G, Becerra-Tomás N, et al. Association between adiposity after diagnosis of prostate cancer and mortality: systematic review and meta-analysis. BMJ Med 2023;2(1):e000339. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20. Rock CL, Thomson CA, Sullivan KR, et al. American Cancer Society nutrition and physical activity guideline for cancer survivors. CA Cancer J Clin. 2022;72(3):230-262. [DOI] [PubMed] [Google Scholar]
  • 21. World Cancer Research Fund International/American Institute for Cancer Research. Diet, Nutrition, Physical activity and cancer: A Global Perspective. Continuous Update Project Report. 2018. https://www.wcrf.org/dietandcancer. Accessed December 10, 2023.
  • 22. Cantarutti A, Bonn SE, Adami HO, et al. Body mass index and mortality in men with prostate cancer. Prostate. 2015;75(11):1129-1136. [DOI] [PubMed] [Google Scholar]
  • 23. Markham MJ, Wachter K, Agarwal N, et al. Clinical cancer advances 2020: annual report on progress against cancer from the American Society of Clinical Oncology. J Clin Oncol. 2020;38(10):1081. [DOI] [PubMed] [Google Scholar]
  • 24. Riboli E, Hunt KJ, Slimani N, et al. European Prospective Investigation into Cancer and Nutrition (EPIC): study populations and data collection. Public Health Nutr. 2002;5(6b):1113-1124. [DOI] [PubMed] [Google Scholar]
  • 25. International Classification of Diseases for Oncology (ICD-O-3). World Health Organization. 2013. https://iris.who.int/handle/10665/42344. Accessed June 20, 2023.
  • 26. Riboli E, Kaaks R.. The EPIC Project: Rationale and study design. European Prospective Investigation into Cancer and Nutrition. Int J Epidemiol. 1997;26(suppl 1):S6-14. [DOI] [PubMed] [Google Scholar]
  • 27. Harrell FE Jr. General aspects of fitting regression models. In: Regression Modeling Strategies: With Applications to Linear Models, Logistic and Ordinal Regression, and Survival Analysis. New York, NY: Springer; 2015:13-44. https://link.springer.com/book/10.1007/978-3-319-19425-7
  • 28. Desquilbet L, Mariotti F.. Dose-response analyses using restricted cubic spline functions in public health research. Stat Med. 2010;29(9):1037-1057. [DOI] [PubMed] [Google Scholar]
  • 29. Harrell FE Jr. Package ‘rms’.2017;229. https://cran.r-project.org/web/packages/rms/index.html. Accessed May 15, 2023.
  • 30. Heinzl H, Kaider A.. Gaining more flexibility in Cox proportional hazards regression models with cubic spline functions. Comput Methods Programs Biomed. 1997;54(3):201-208. [DOI] [PubMed] [Google Scholar]
  • 31. Ferrari P, Day NE, Boshuizen HC, et al. The evaluation of the diet/disease relation in the EPIC study: considerations for the calibration and the disease models. Int J Epidemiol. 2008;37(2):368-378. [DOI] [PubMed] [Google Scholar]
  • 32. Hughes RA, Heron J, Sterne JAC, Tilling K.. Accounting for missing data in statistical analyses: multiple imputation is not always the answer. Int J Epidemiol. 2019;48(4):1294-1304. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33. Vach W. Some issues in estimating the effect of prognostic factors from incomplete covariate data. Statist Med 1997;16(1):57-72. [DOI] [PubMed] [Google Scholar]
  • 34. Schoenfeld D. Partial residuals for the proportional hazards regression model. Biometrika. 1982;69(1):239-241. [Google Scholar]
  • 35. Grambsch PM, Therneau TM.. Proportional hazards tests and diagnostics based on weighted residuals. J Biometrika. 1994;81(3):515-526. [Google Scholar]
  • 36. Hess KR. Graphical methods for assessing violations of the proportional hazards assumption in Cox regression. Stat Med. 1995;14(15):1707-1723. [DOI] [PubMed] [Google Scholar]
  • 37. Wang LS, Murphy CT, Ruth K, et al. Impact of obesity on outcomes after definitive dose-escalated intensity-modulated radiotherapy for localized prostate cancer. Cancer. 2015;121(17):3010-3017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38. Xu MC, Huelster HL, Hatcher JB, et al. Obesity is associated with longer survival independent of sarcopenia and myosteatosis in metastatic and/or castrate-resistant prostate cancer. J Urol. 2020;205(3):800-805. doi: 10.1097/ju.0000000000001428. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39. Vidal AC, Howard LE, de Hoedt A, et al. Obese patients with castration-resistant prostate cancer may be at a lower risk of all-cause mortality: Results from the Shared Equal Access Regional Cancer Hospital (SEARCH) database. BJU Int. 2018;122(1):76-82. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40. Halabi S, Ou SS, Vogelzang NJ, Small EJ.. Inverse correlation between body mass index and clinical outcomes in men with advanced castration-recurrent prostate cancer. Cancer. 2007;110(7):1478-1484. [DOI] [PubMed] [Google Scholar]
  • 41. Martini A, Shah QN, Waingankar N, et al. The obesity paradox in metastatic castration-resistant prostate cancer. Prostate Cancer Prostatic Dis. 2021;25(3):472-478. doi: 10.1038/s41391-021-00418-0. [DOI] [PubMed] [Google Scholar]
  • 42. Cao Y, Ma J.. Body mass index, prostate cancer-specific mortality, and biochemical recurrence: a systematic review and meta-analysis. Cancer Prev Res (Phila). 2011;4(4):486-501. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43. Zhong S, Yan X, Wu Y, et al. Body mass index and mortality in prostate cancer patients: a dose-response meta-analysis. Prostate Cancer Prostatic Dis. 2016;19(2):122-131. [DOI] [PubMed] [Google Scholar]
  • 44. Rivera-Izquierdo M, Pérez de Rojas J, Martínez-Ruiz V, et al. Obesity as a risk factor for prostate cancer mortality: a systematic review and dose-response meta-analysis of 280,199 patients. Cancers (Basel). 2021;13(16):4169. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45. Lavalette C, Cordina Duverger E, Artaud F, et al. Body mass index trajectories and prostate cancer risk: results from the EPICAP study. Cancer Med. 2020;9(17):6421-6429. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46. Kelly SP, Lennon H, Sperrin M, et al. Body mass index trajectories across adulthood and smoking in relation to prostate cancer risks: the NIH-AARP Diet and Health Study. Int J Epidemiol. 2019;48(2):464-473. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47. Dickerman BA, Ahearn TU, Giovannucci E, et al. Weight change, obesity and risk of prostate cancer progression among men with clinically localized prostate cancer. Int J Cancer. 2017;141(5):933-944. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48. Song M, Willett WC, Hu FB, et al. Trajectory of body shape across the lifespan and cancer risk. Int J Cancer. 2016;138(10):2383-2395. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49. Jackson MD, Tulloch-Reid MK, McCaw-Binns AM, et al. Central adiposity at diagnosis may reduce prostate cancer-specific mortality in African-Caribbean men with prostate cancer: 10-year follow-up of participants in a case-control study. Cancer Causes Control. 2020;31(7):651-662. [DOI] [PubMed] [Google Scholar]
  • 50. Farris MS, Courneya KS, Kopciuk KA, McGregor SE, Friedenreich CM.. Anthropometric measurements and survival after a prostate cancer diagnosis. Br J Cancer. 2018;118(4):607-610. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51. Polesel J, Gini A, Dal Maso L, et al. The impact of diabetes and other metabolic disorders on prostate cancer prognosis. J Diabetes Complications. 2016;30(4):591-596. [DOI] [PubMed] [Google Scholar]
  • 52. Wilson RL, Taaffe DR, Newton RU, et al. Obesity and prostate cancer: A narrative review. Crit Rev Oncol Hematol. 2022;169:103543. [DOI] [PubMed] [Google Scholar]
  • 53. Chau CH, Till C, Price DK, et al. Serum markers, obesity and prostate cancer risk: Results from the prostate cancer prevention trial. Endocr Relat Cancer. 2022;29(2):99-109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54. Pham DV, Nguyen TK, Park PH.. Adipokines at the crossroads of obesity and mesenchymal stem cell therapy. Exp Mol Med. 2023;55(2):313-324. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55. Di Francesco S, Robuffo I, Caruso M, et al. Metabolic alterations, aggressive hormone-naïve prostate cancer and cardiovascular disease: a complex relationship. Medicina (Kaunas). 2019;55(3):62. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56. Arcidiacono B, Iiritano S, Nocera A, et al. Insulin resistance and cancer risk: an overview of the pathogenetic mechanisms. Exp Diabetes Res. 2012;2012:789174. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57. Cho HJ, Kwon GT, Park H, et al. A high-fat diet containing lard accelerates prostate cancer progression and reduces survival rate in mice: possible contribution of adipose tissue-derived cytokines. Nutrients. 2015;7(4):2539-2561. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 58. Stark T, Livas L, Kyprianou N.. Inflammation in prostate cancer progression and therapeutic targeting. Transl Androl Urol. 2015;4(4):455-463. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 59. Archer M, Dogra N, Kyprianou N.. Inflammation as a driver of prostate cancer metastasis and therapeutic resistance. Cancers. 2020;12(10):2984. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 60. Kaaks R, Stattin P.. Obesity, endogenous hormone metabolism, and prostate cancer risk: a conundrum of "highs" and "lows". Cancer Prev Res (Phila). 2010;3(3):259-262. [DOI] [PubMed] [Google Scholar]
  • 61. Ormazabal V, Nair S, Elfeky O, et al. Association between insulin resistance and the development of cardiovascular disease. Cardiovasc Diabetol. 2018;17(1):122. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 62. Rodríguez-Rodríguez E, López-Sobaler AM, Ortega RM, et al. Association between neutrophil-to-lymphocyte ratio with abdominal obesity and healthy eating index in a representative older Spanish population. Nutrients. 2020;12(3):855. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 63. Uribe-Querol E, Rosales C.. Neutrophils actively contribute to obesity-associated inflammation and pathological complications. Cells. 2022;11(12):1883. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 64. Cupp MA, Cariolou M, Tzoulaki I, et al. Neutrophil to lymphocyte ratio and cancer prognosis: An umbrella review of systematic reviews and meta-analyses of observational studies. BMC Med. 2020;18(1):360. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 65. Champ CE, Francis L, Klement RJ, Dickerman R, Smith RP.. Fortifying the Treatment of Prostate Cancer with Physical Activity. Prostate Cancer. 2016;2016:9462975. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 66. Schmidt DR, Patel R, Kirsch DG, et al. Metabolomics in cancer research and emerging applications in clinical oncology. CA Cancer J Clin. 2021;71(4):333-358. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 67. Allott EH, Hursting SD.. Obesity and cancer: mechanistic insights from transdisciplinary studies. Endocr Relat Cancer. 2015;22(6):R365-86. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 68. Vidal AC, Howard LE, Sun SX, et al. Obesity and prostate cancer-specific mortality after radical prostatectomy: results from the Shared Equal Access Regional Cancer Hospital (SEARCH) database. Prostate Cancer Prostatic Dis. 2017;20(1):72-78. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 69. Keto CJ, Aronson WJ, Terris MK, et al. Obesity is associated with castration-resistant disease and metastasis in men treated with androgen deprivation therapy after radical prostatectomy: results from the SEARCH database. BJU Int. 2012;110(4):492-498. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 70. Mangiola S, Stuchbery R, McCoy P, et al. Androgen deprivation therapy promotes an obesity-like microenvironment in periprostatic fat. Endocr Connect. 2019;8(5):547-558. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 71. Tonorezos ES, Jones LW.. Energy balance and metabolism after cancer treatment. Semin Oncol. 2013;40(6):745-756. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 72. Manthri S, Geraci SA, Chakraborty K.. Overview of cancer survivorship care for primary care providers. Cureus. 2020;12(9):e10210. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 73. Bonet C, Crous-Bou M, Tsilidis KK, et al. The association between body fatness and mortality among breast cancer survivors: Results from a prospective cohort study. Eur J Epidemiol. 2023;38(5):545-557. doi: 10.1007/s10654-023-00979-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 74. Bonn SE, Wiklund F, Sjölander A, et al. Body mass index and weight change in men with prostate cancer: progression and mortality. Cancer Causes Control. 2014;25(8):933-943. [DOI] [PubMed] [Google Scholar]
  • 75. Parker C, Castro E, Fizazi K, et al. ; ESMO Guidelines Committee. Electronic address: Clinicalguidelines@esmo.org. Prostate cancer: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2020;31(9):1119-1134. [DOI] [PubMed] [Google Scholar]
  • 76. Ruiz de Porras V, Font A, Aytes A.. Chemotherapy in metastatic castration-resistant prostate cancer: Current scenario and future perspectives. Cancer Lett. 2021;523:162-169. [DOI] [PubMed] [Google Scholar]
  • 77. Kim MM, Hoffman KE, Levy LB, et al. Improvement in prostate cancer survival over time: A 20-year analysis. Cancer J. 2012;18(1):1-8. [DOI] [PubMed] [Google Scholar]
  • 78. Nevedomskaya E, Baumgart SJ, Haendler B.. Recent advances in prostate cancer treatment and drug discovery. Int J Mol Sci. 2018;19(5): [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 79. Jones MP. Indicator and stratification methods for missing explanatory variables in multiple linear regression. J Am Stat Assoc. 1996;91(433):222-230. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

pkae070_Supplementary_Data
pkae070_supplementary_data.pdf (2.4MB, pdf)

Data Availability Statement

For information about how to apply for gaining access to EPIC data or biospecimens, please follow the instructions at https://epic.iarc.fr/access.

Articles from JNCI Cancer Spectrum are provided here courtesy of Oxford University Press

RESOURCES