/ Men (n=1252) / Hi2b / Women (n=1268) / Hi2b /
/ Component 1 / Component 2 / Component 3 / Component 4 / Component 1 / Component 2 / Component 3 / Component 4 /
HDL-cholesterol / 64* / -2 / -9 / -34 / 0.53 / 61 / -40 / 1 / 5 / 0.54
HMW-adiponectin / 84* / 2 / -6 / 0 / 0.71 / 82 / -13 / -7 / 14 / 0.72
Non-HMW adiponectin / 83* / -10 / -2 / 5 / 0.69 / 82 / -1 / -7 / 6 / 0.68
ROM / 3 / 77* / -4 / 6 / 0.61 / -2 / -1 / 77 / -8 / 0.60
CRP / -5 / 80* / 16 / -4 / 0.67 / -14 / 5 / 75 / -14 / 0.61
Triglycerides / -29 / -3 / 13 / 74* / 0.65 / -16 / 79 / -2 / -5 / 0.65
C-peptide / -3 / 5 / 42 / 55* / 0.48 / -4 / 64 / 10 / -38 / 0.56
Leptin / 8 / 13 / 78* / 13 / 0.65 / -6 / 27 / 15 / -68 / 0.56
sOB-R / 25 / 5 / -80* / 4 / 0.70 / 15 / 8 / 2 / 81 / 0.68
Hba1c / -9 / 46 / -2 / 38 / 0.36 / -16 / 47 / 38 / 20 / 0.43
IGF1 / -31 / -19 / 18 / -51 / 0.42 / -33 / -27 / -45 / -19 / 0.42
Total variance explained (%) / 22.3 / 15.4 / 11.6 / 9.6 / 24.4 / 14.4 / 10.5 / 9.5
Cumulative total variance (%) / 22.3 / 37.6 / 49.3 / 58.8 / 24.4 / 38.8 / 49.2 / 58.8
aThe principal axis method was used to extract the components, and this was followed by a varimax (orthogonal) rotation. Printed values are multiplied by 100 and rounded to the nearest integer. Values with factor loadings greater than 0.55 are flagged by an '*'.
bCommunality estimates (proportion of variance of the specific biomarker explained by the four components) appear in column headed hi2.
Note: Analysis in men is based on 626 cases/626 controls and in women on 634 cases and 634 controls.
Abbreviations: HbA1c=glycated hemoglobin; IGF-1= insulin growth factor 1; HDL-cholesterol = high-density-lipoprotein cholesterol; CRP = C-reactive protein; ROM = reactive oxygen metabolites; HMW-adiponectin = high-molecular-weight adiponectin; sOB-R = soluble leptin receptor.
Supplementary Table 2. Incidence Rate Ratios and 95% Confidence of Biomarker Patternsa with Colorectal, Colon and Rectal Cancer in Men, the European Prospective Investigation into Cancer and Nutrition (1992-2003)Biomarker pattern / Q1 / Q2 / Q3 / Q4 / P for trendc
Component 1 (HDL-C/Adiponectin)
Cases/Controls / 200/157 / 154/157 / 129/157 / 142/157
Crude model / 1 / 0.76 (0.54-1.06) / 0.62 (0.45-0.87) / 0.67 (0.48-0.94) / 0.01
Multivariable modelb / 1 / 0.75 (0.53-1.06) / 0.66 (0.46-0.93) / 0.72 (0.50-1.02) / 0.06
Component 2 (ROM/CRP)
Cases/Controls / 126/157 / 126/157 / 170/157 / 204/157
Crude model / 1 / 0.99 (0.66-1.48) / 1.48 (1.03-2.13) / 1.72 (1.32-2.43) / 0.0002
Multivariable modelb / 1 / 0.91 (0.61-1.40) / 1.34 (0.93-1.94) / 1.50 (1.03-2.18) / 0.003
Component 3 (leptin/sOB-R)
Cases/Controls / 150/157 / 138/157 / 164/157 / 173/157
Crude model / 1 / 0.92 (0.62-1.36) / 1.10 (0.77-1.56) / 1.18 (0.81-1.72) / 0.25
Multivariable modelb / 1 / 0.88 (0.58-1.32) / 0.98 (0.66-1.45) / 0.94 (0.59-1.47) / 0.88
Component 4 (TG/C-peptide)
Cases/Controls / 159/157 / 159/157 / 148/157 / 161/157
Crude model / 1 / 1.01 (0.69-1.48) / 0.93 (0.66-1.32) / 1.01 (0.71-1.45) / 0.99
Multivariable modelb / 1 / 1.01 (0.67-1.51) / 0.85 (0.59-1.22) / 0.85 (0.58-1.25) / 0.32
aBiomarkers found to load on the first component were adiponectin and HDL-cholesterol; biomarkers mostly loading on the second component were triglycerides and C-peptide, biomarkers mostly loading the third component were ROM and CRP; and biomarkers loading on the fourth components were leptin and sOB-R.
bMultivariable model was based on the crude model taking into account matching factors: age, sex, study centre, follow-up time since blood collection, time of the day at blood collection and fasting status; Women were further matched by menopausal status, phase of menstrual cycle at blood collection and postmenopausal women were matched by HRT use; with additional adjustment for education (no school degree or primary school, technical or professional school, secondary school, university degree, or unknown), smoking status (never, past, current, or unknown), physical activity (inactive, moderately inactive, moderately active, active, or missing), alcohol intake (continuous), fruits and vegetables (g/day), fish and shellfish intake (g/day), red and processed meat (g/day), fibre intake (g/day) and waist circumference (cm).
cP-value for trend (two sided) across quartiles is based on the median principal component score within quartiles as a continuous variable. Wald χ2 test was employed to evaluate the significance of the linear trend.
Abbreviations: HDL-C =high-density-lipoprotein cholesterol; CRP = high-sensitivity C-reactive protein; TG = triglycerides; ROM = reactive oxygen metabolites; sOB-R = soluble leptin receptor; All P values are two-sided.
Supplementary Table 3. Incidence Rate Ratios and 95% Confidence of Biomarker Patternsa with Colorectal Cancer in Women, the European Prospective Investigation into Cancer and Nutrition (1992-2003)Biomarker pattern / Q1 / Q2 / Q3 / Q4 / P for trendc
Component 1 (HDL-C/Adiponectin)
Cases/Controls / 193/158 / 140/159 / 157/158 / 144/158
Crude model / 1 / 0.70 (0.48-1.03) / 0.73 (0.51-1.06) / 0.68 (0.48-0.96) / 0.10
Multivariable modelb / 1 / 0.72 (0.48-1.07) / 0.74 (0.50-1.08) / 0.73 (0.50-1.05) / 0.12
Component 2 (TG/C-peptide )
Cases/Controls / 181/158 / 146159 / 131/158 / 176/159
Crude model / 1 / 0.80 (0.56-1.15) / 0.71 (0.48-1.06) / 0.96 (0.67-1.39) / 0.89
Multivariable modelb / 1 / 0.78 (0.54-1.13) / 0.67 (0.44-1.03) / 0.84 (0.56-1.25) / 0.37
Component 3 (ROM/CRP)
Cases/Controls / 133/159 / 137/158 / 170/159 / 194/159
Crude model / 1 / 1.07 (0.70-1.64) / 1.34 (0.87-2.06) / 1.56 (1.04-2.32) / 0.01
Multivariable modelb / 1 / 1.04 (0.67-1.61) / 1.27 (0.82-1.97) / 1.45 (0.95-2.21) / 0.04
Component 4 (leptin/sOB-R)
Cases/Controls / 167/159 / 163/158 / 187/159 / 117/159
Crude model / 1 / 0.97 (0.69-1.27) / 1.13 (0.81-1.58) / 0.70 (0.48-1.01) / 0.12
Multivariable modelb / 1 / 1.08 (0.75-1.56) / 1.32 (0.91-1.90) / 0.84 (0.55-1.28) / 0.61
aBiomarkers found to load on the first component were adiponectin and HDL-cholesterol; biomarkers mostly loading on the second component were ROM and CRP, biomarkers mostly loading the third component were leptin and sOB-R; and biomarkers loading on the fourth components were triglycerides and C-peptide.
bMultivariable model was based on the crude model taking into account matching factors: age, sex, study centre, follow-up time since blood collection, time of the day at blood collection and fasting status; Women were further matched by menopausal status, phase of menstrual cycle at blood collection and postmenopausal women were matched by HRT use; with additional adjustment for education (no school degree or primary school, technical or professional school, secondary school, university degree, or unknown), smoking status (never, past, current, or unknown), physical activity (inactive, moderately inactive, moderately active, active, or missing), alcohol intake (continuous), fruits and vegetables (g/day), fish and shellfish intake (g/day), red and processed meat (g/day), fibre intake (g/day) and waist circumference (cm).
cP-value for trend (two sided) across quartiles is based on the median principal component score within quartiles as a continuous variable. Wald χ2 test was employed to evaluate the significance of the linear trend.
Abbreviations: HDL-C =high-density-lipoprotein cholesterol; CRP = high-sensitivity C-reactive protein; TG = triglycerides; ROM = reactive oxygen metabolites; sOB-R = soluble leptin receptor; All P values are two-sided.
Supplementary table 4. Incidence Rate Ratios and 95% Confidence Intervals of Biomarker Patternsa with Colorectal Cancer by tumor stage*, the European Prospective Investigation into Cancer and Nutrition (1992-2003)Biomarker pattern / Q1 / Q2 / Q3 / Q4 / P for trendd
Component 1
(HDL-C/Adiponectin fractions)
Stage I and II / 1 / 0.97 (0.64-1.46) / 0.95 (0.62-1.47) / 0.84 (0.54-1.37) / 0.05
Stage III and IV / 1 / 0.53 (0.32-0.84) / 0.52 (0.32-0.85) / 0.50 (0.29-0.86) / 0.01
Component 2
(ROM/CRP)
Stage I and II / 1 / 1.11 (0.84-1.47) / 1.17 (0.89-1.55) / 1.56 (1.16-2.09) / 0.01
Stage III and IV / 1 / 1.06 (0.68-1.66) / 1.26 (0.77-2.04) / 1.43 (0.92-2.23) / 0.08
Component 3
(TG/C-peptide)
Stage I and II / 1 / 0.92 (0.59-1.41) / 0.72 (0.45-1.14) / 0.71 (0.43-1.16) / 0.13
Stage III and IV / 1 / 0.71 (0.46-1.11) / 1.11 (0.71-1.76) / 0.81 (0.49-1.34) / 0.75
Component 4
(leptin/sOB-R)
Stage I and II / 1 / 0.96 (0.61-1.48) / 0.64 (0.38-1.04) / 0.74 (0.45-1.20) / 0.10
Stage III and IV / 1 / 1.19 (0.78-1.83) / 1.17 (0.73-1.87) / 1.11 (0.67-1.85) / 0.66
aBiomarkers found to load on the first component were adiponectin and HDL-cholesterol; biomarkers mostly loading on the second component were triglycerides and C-peptide, biomarkers mostly loading the third component were ROM and CRP; and biomarkers loading on the fourth component were leptin and sOB-R.
bMultivariable model was based on the crude model taking into account matching factors: age, sex, study centre, follow-up time since blood collection, time of the day at blood collection and fasting status; Women were further matched by menopausal status, phase of menstrual cycle at blood collection and postmenopausal women were matched by HRT use; with additional adjustment for education (no school degree or primary school, technical or professional school, secondary school, university degree, or unknown), smoking status (never, past, current, or unknown), physical activity (inactive, moderately inactive, moderately active, active, or missing), alcohol intake (continuous), fruits and vegetables (g/day), fish and shellfish intake (g/day), red and processed meat (g/day), fibre intake (g/day) and waist circumference (cm).
cP-value for trend (two sided) across quartiles is based on the median principal component score within quartiles as a continuous variable. Wald χ2 test was employed to evaluate the significance of the linear trend.
dAnalysis based on 1129 cases and 1129 controls with available information on tumor stage.
The distribution of CRC cases by tumor stage was as follows: Stage I (localized), n=230; 20.37%; Stage II (localized with invasion), n=215; 19.04%; Stage III = metastatic regional; n=339 (30.03%); Stage IV; metastatic distal, n=120 (10.63%); Unknown, N=220 (16.5%).
Abbreviations: HDL-C =high-density-lipoprotein cholesterol; CRP = high-sensitivity C-reactive protein; TG = triglycerides; ROM = reactive oxygen metabolites; sOB-R = soluble leptin receptor; All P values are two-sided.
Supplementary Table 5: Summary of the scientific research findings on individual lifestyle and metabolic factors and colorectal cancer incidence in the EPIC StudyBiomarker / Main underlying mechanisms / Association with colon cancer / RR (95% CI)
C-peptidea / Hyperinsulinemia / for the highest versus the lowest quintile / 1.67 (1.14-2.46)
P-trend<0.01
Glycated hemoglobin (HbA1c) b / Hyperglycemia / for a 10% increase in HbA1c / 1.10 (1.01-1.19 )
High-density lipoprotein cholesterol
(HDL-C)c / Hyperalphalipoproteinemia / for 1 SD increase of 16.6 mg/dl in HDL / 0.78 (0.68 -0.89)
P-trend < 0.0001
High-sensitivity
C-reactive protein (CRP-hs) d / Chronic low grade Inflammation / for CRP concentrations of > or = 3.0 mg/L versus <1.0 mg/L / 1.36 (1.00-1.85)
P-trend = 0.01
Reactive oxygen metabolites (ROM)e / Oxidative stress / comparing the highest tertile vs. the lowest / 1.91(1.47-2.48)
P-trend < 0.0001
Non-high molecular weight (HMW) adiponectinf / Alterations via adipokine overproduction / comparing the highest quintile vs. the lowest / 0.45 (0.34-0.61)
P-trend < 0.0001
Soluble Leptin receptor (sOB-R)g / comparing the highest quintile vs. the lowest / 0.42 (0.28-0.63)
P-trend = 0.0001
aJenab M, International journal of cancer. 2007;121(2):368-376.
bRinaldi S, American Association for Cancer Research 2008;17(11):3108-3115.
cvan Duijnhoven FJ. Gut. Aug 2011;60(8):1094-1102.
dAleksandrova K. American journal of epidemiology. 2010;172(4):407-418.
eLeufkens AM. American journal of epidemiology.175(7):653-663.
fAleksandrova K. Cancer research. 2012;72(20):5328-5337.
gAleksandrova K. Carcinogenesis. 2012;33(6):1211-1218.