DNA methylation ageing clocks and pancreatic cancer risk: pooled analysis of three prospective nested case-control studies

Figures & data

Table 1. Baseline characteristics of the study population, stratified by epigenetic age acceleration or deceleration^a.

Mean (SD) or n (%)	Overall	AA Hannum		AA Horvath		AA PhenoAge
Mean (SD) or n (%)	Overall	Deceleration	Acceleration	Deceleration	Acceleration	Deceleration	Acceleration
n	824	407	407	405	411	399	419
Pancreatic cancer cases^b	393 (47.7%)	182 (44.7%)	203 (49.9%)	184 (45.4%)	204 (49.6%)	178 (44.6%)	211 (50.4%)
Cohort
NHS	370 (44.9%)	210 (51.6%)	160 (39.3%)	203 (50.1%)	166 (40.4%)	186 (46.6%)	184 (43.9%)
HPFS	297 (36.0%)	128 (31.4%)	159 (39.1%)	134 (33.1%)	157 (38.2%)	137 (34.3%)	154 (36.8%)
PHS	157 (19.1%)	69 (17.0%)	88 (21.6%)	68 (16.8%)	88 (21.4%)	76 (19.0%)	81 (19.3%)
Age at blood draw, y	60.4 (7.8)	59.3 (7.9)	61.3 (7.5)	59.5 (8.0)	61.1 (7.5)	59.7 (8.0)	61.0 (7.6)
Time before diagnosis, y^c	13.0 (6.2)	13.1 (6.4)	12.9 (6.1)	13.1 (6.5)	12.9 (6.0)	13.1 (6.3)	12.8 (6.2)
Female	370 (44.9%)	210 (51.6%)	160 (39.3%)	203 (50.1%)	166 (40.4%)	186 (46.6%)	184 (43.9%)
White^d	752 (94.5%)	365 (92.9%)	377 (95.9%)	365 (92.6%)	380 (96.2%)	367 (94.8%)	379 (94.0%)
Smoking status^e
Never	332 (40.6%)	159 (39.3%)	169 (41.8%)	158 (39.4%)	171 (41.7%)	166 (42.0%)	164 (39.3%)
Former	364 (44.5%)	181 (44.7%)	178 (44.1%)	179 (44.6%)	181 (44.1%)	171 (43.3%)	189 (45.3%)
Current	122 (14.9%)	65 (16.0%)	57 (14.1%)	64 (16.0%)	58 (14.1%)	58 (14.7%)	64 (15.3%)
BMI^f, kg/m^2	25.8 (4.1)	26.0 (4.2)	25.7 (4.0)	25.7 (4.1)	26.0 (4.1)	25.6 (3.9)	26.1 (4.3)
Normal and lower (BMI <25)	398 (49.1%)	197 (48.9%)	197 (49.6%)	205 (51.5%)	191 (47.3%)	205 (51.9%)	190 (46.5%)
Overweight (BMI 25 to <30)	295 (36.4%)	146 (36.2%)	143 (36.0%)	139 (34.9%)	150 (37.1%)	143 (36.2%)	149 (36.4%)
Obese (BMI ≥30)	117 (14.4%)	60 (14.9%)	57 (14.4%)	54 (13.6%)	63 (15.6%)	47 (11.9%)	70 (17.1%)
Diabetes^g	30 (3.6%)	13 (3.2%)	17 (4.2%)	15 (3.7%)	15 (3.6%)	12 (3.0%)	18 (4.3%)

AA: epigenetic age acceleration metric; BMI: body mass index; HPFS: Health Professionals Follow-up Study; IEAA: intrinsic epigenetic age acceleration metric; n: sample size; NHS: Nurses’ Health Study; PHS: Physicians’ Health Study; SD: standard deviation.

^aSubjects with the absolute value of the age acceleration estimate greater than three SDs were excluded.

^bCase status based on the end of the cohort (other covariates based on questionnaires closest to time of blood draw).

^c11 missing values.

^d28 missing values.

^e6 missing values.

^f14 missing values.

^g1 missing values.

Figure 1. Forest plot of the associations between age acceleration metrics (per quintile) and pancreatic cancer risk

AA = epigenetic age acceleration metric; CI = confidence interval; IEAA = intrinsic epigenetic age acceleration metric; ref = reference group. Please refer to Table footnotes for quartile cut-offs.

Figure 2. Results from the spline regression analyses of the associations between each IEAA metric and pancreatic cancer risks (panel A, IEAA Hannum; panel B, IEAA Horvath; panel C, IEAA PhenoAge)

IEAA = intrinsic epigenetic age acceleration. All models adjusted for age at blood draw, sex, race, date of blood draw, smoking status, and BMI. Solid curve line = smoothing spline-fitted estimate of pancreatic cancer risk (log scale). Dotted line = standard error of the fitted estimate. Grey areas indicate elevated pancreatic risks.

Figure 3. Stratified spline regression analyses of the associations between each IEAA metric and pancreatic cancer risks by smoking status (panel A, Never smokers; panel B, Former smokers; panel C, Current smokers)

IEAA = intrinsic epigenetic age acceleration. All models adjusted for age at blood draw, sex, race, date of blood draw, smoking status, and BMI. Solid curve line = smoothing spline-fitted estimate of pancreatic cancer risk (log scale). Dotted line = standard error of the fitted estimate.

Figure 4. Stratified spline regression analyses of the associations between each IEAA metric and pancreatic cancer risks by BMI groups (panel A, BMI ≤ 25 kg/m²; panel B, BMI > 25 kg/m²)

IEAA = intrinsic epigenetic age acceleration. All models adjusted for age at blood draw, sex, race, date of blood draw, smoking status, and BMI. Solid curve line = smoothing spline-fitted estimate of pancreatic cancer risk (log scale). Dotted line = standard error of the fitted estimate.

Table 2. Associations between epigenetic age acceleration metrics (higher quintile vs. the lowest quintile) and pancreatic cancer survival in pancreatic cancer cases

Epigenetic age acceleration metrics	Quartiles^a	N	HR (95% CI)^b	P-value
AA Hannum	Q1	69	1 (ref)
	Q2	83	1.36 (0.97, 1.92)	0.08
	Q3	100	1.24 (0.89, 1.73)	0.21
	Q4	82	1.07 (0.77, 1.50)	0.68
AA Horvath	Q1	61	1 (ref)
	Q2	79	1.21 (0.87, 1.68)	0.26
	Q3	95	1.17 (0.86, 1.61)	0.32
	Q4	97	0.94 (0.67, 1.32)	0.72
AA PhenoAge	Q1	63	1 (ref)
	Q2	84	1.21 (0.86, 1.69)	0.27
	Q3	97	1.10 (0.80, 1.52)	0.56
	Q4	91	1.30 (0.92, 1.82)	0.13
IEAA Hannum	Q1	62	1 (ref)
	Q2	83	1.45 (1.02, 2.05)	0.04
	Q3	115	1.29 (0.91, 1.83)	0.16
	Q4	74	1.13 (0.79, 1.61)	0.5
IEAA Horvath	Q1	53	1 (ref)
	Q2	95	1.19 (0.85, 1.67)	0.3
	Q3	96	1.13 (0.82, 1.56)	0.46
	Q4	88	1.09 (0.76, 1.56)	0.64
IEAA PhenoAge	Q1	63	1 (ref)
	Q2	72	1.15 (0.81, 1.63)	0.44
	Q3	110	1.16 (0.84, 1.59)	0.37
	Q4	89	1.10 (0.78, 1.54)	0.58

AA: epigenetic age acceleration metric; CI: confidence interval; IEAA: intrinsic epigenetic age acceleration metric; n: sample size; HR: hazard ratio; ref: reference group.

^aQuartile (Q) cut-offs: Q1: <−3.2 for AA Hannum, <−3.7 for AA Horvath, <−4.8 for AA PhenoAge, <−3.2 for IEAA Hannum, <−3.8 for IEAA Horvath, and <−4.2 for IEAA PhenoAge; Q2: −3.2 to −0.31 for AA Hannum, −3.7 to −0.21 for AA Horvath, −4.8 to −0.41 for AA PhenoAge, −3.2 to −0.11 for IEAA Hannum, −3.8 to −0.11 for IEAA Horvath, and −4.2 to −0.61 for IEAA PhenoAge; Q3: −0.3 to 3.1 for AA Hannum, −0.2 to 3.9 for AA Horvath, −0.4 to 4.2 for AA PhenoAge, −0.1 to 3.3 for IEAA Hannum, −0.1 to 3.9 for IEAA Horvath, and −0.6 to 3.9 for IEAA PhenoAge; Q4: ≥3.1for AA Hannum, ≥3.9 for AA Horvath, ≥4.2 for AA_PhenoAge, ≥3.3 for IEAA Hannum, ≥3.9 for IEAA Horvath, and ≥ 3.9 for IEAA PhenoAge.

^bCox proportional hazard regression models were adjusted for cohorts, age at blood draw, race, date of blood draw, smoking status, and BMI.

Data availability statement

All data from this study have been deposited in dbGAP [‘DNA Methylation Markers and Pancreatic Cancer Risk in 3 Cohort Studies (NHS, PHS, HPFS)’ phs001917.v1.p1]. https://www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/study.cgi?study_id=phs001917.v1.p1