Table 2.

OR of pancreatic cancer by plasma quartiles of folate, vitamin B6, vitamin B12, and homocysteine in four prospective cohorts combined

Plasma one-carbon nutrientQ1Q2Q3Q4Ptrend
Main model
Folate (ng/mL) *
    Cases/controls 53/16547/16847/12459/157
    Simple model1.00.90 (0.57–1.43)1.23 (0.76–2.00)1.20 (0.76–1.91)0.30
    Multivariable model1.00.89 (0.56–1.42)1.26 (0.77–2.07)1.22 (0.77–1.95)0.28
    Mutual adjustment1.00.85 (0.53–1.37)1.20 (0.71–2.04)1.34 (0.79–2.26)0.11
Vitamin B6 (pmol/mL) *
    Cases/controls 58/15441/13259/17249/162
    Simple model1.00.82 (0.51–1.31)0.93 (0.61–1.42)0.80 (0.51–1.25)0.25
    Multivariable model1.00.88 (0.55–1.42)1.01 (0.66–1.55)0.87 (0.55–1.37)0.37
    Mutual adjustment1.00.81 (0.49–1.33)0.93 (0.58–1.47)0.83 (0.49–1.39)0.29
Vitamin B12 (pg/mL) *
    Cases/controls 47/15455/15564/15642/155
    Simple model1.01.16 (0.74–1.81)1.34 (0.87–2.07)0.91 (0.57–1.46)0.71
    Multivariable model1.01.17 (0.74–1.83)1.42 (0.92–2.21)0.93 (0.58–1.49)0.79
    Mutual adjustment1.01.23 (0.76–1.99)1.40 (0.87–2.28)0.86 (0.50–1.46)0.68
Homocysteine (μmol/L) *
    Cases/controls 40/16955/13252/13860/182
    Simple model1.01.73 (1.09–2.74)1.61 (1.00–2.61)1.43 (0.90–2.28)0.24
    Multivariable model1.01.63 (1.02–2.62)1.54 (0.94–2.52)1.37 (0.85–2.20)0.35
    Mutual adjustment1.01.73 (1.05–2.84)1.57 (0.94–2.61)1.48 (0.89–2.48)0.38
Excluding multivitamin users
Folate (ng/mL) *
    Cases/controls 20/5232/10227/10839/130
    Simple model1.00.58 (0.26–1.29)0.45 (0.19–1.03)0.55 (0.24–1.25)0.19
    Multivariable model1.00.56 (0.24–1.32)0.46 (0.19–1.12)0.54 (0.23–1.30)0.21
    Mutual adjustment1.00.58 (0.24–1.43)0.48 (0.18–1.22)0.63 (0.24–1.62)0.34
Vitamin B6 (pmol/mL) *
Cases/Control 36/8228/9431/11423/104
    Simple model1.00.58 (0.30–1.15)0.59 (0.31–1.10)0.47 (0.24–0.92)0.08
    Multivariable model1.00.65 (0.32–1.32)0.69 (0.36–1.32)0.51 (0.25–1.02)0.15
    Mutual adjustment1.00.78 (0.37–1.65)0.74 (0.37–1.48)0.60 (0.27–3.24)0.43
Vitamin B12 (pg/mL) *
    Cases/controls 31/9737/9627/9924/102
    Simple model1.01.47 (0.78–2.76)1.00 (0.52–1.93)0.80 (0.41–1.54)0.27
    Multivariable model1.01.34 (0.69–2.60)1.01 (0.51–2.02)0.81 (0.41–1.59)0.35
    Mutual adjustment1.01.58 (0.77–3.24)1.23 (0.58–2.57)1.04 (0.49–2.19)0.64
Homocysteine (μmol/L) *
    Cases/controls 16/6729/10532/9742/126
    Simple model1.01.24 (0.60–2.57)1.58 (0.72–3.46)1.42 (0.68–2.94)0.47
    Multivariable model1.01.21 (0.57–2.57)1.59 (0.71–3.55)1.33 (0.63–2.82)0.57
    Mutual adjustment1.01.20 (0.54–2.64)1.62 (0.70–3.76)1.32 (0.59–2.93)0.84
  • NOTE: Quartiles/test for trend (created based on median per quartile) based on controls only; Q1 denotes lowest quartile.

  • * Analyses based on simple conditional logistic regression models adjusting for the matching factors [year of birth, smoking status (current, past, or never), fasting status, month of blood draw] and cohort (i.e., thus also for gender); multivariable models are additionally adjusted for BMI, physical activity, and a history of diabetes; models labeled “mutual adjustment” are further adjusted for the other three biochemical indicators of methyl-group availability.

  • Numbers vary slightly by assay because of missing values and different quartile distributions.

  • Quartiles are based on controls without any multivitamin use only and calculated as cohort specific; thus, ranges of quartiles are not provided because they overlap in analyses combining all four cohorts.