Skip to main content

Table 3 Multivariatea model of the association of polyphenol intakes with TBARs according to the method of handling missing values

From: Association between dietary polyphenols intake and an oxidative stress biomarker: interest of multiple imputation for handling missing covariates and outcomes

Final models

 

Complete case analysis

  

Multiple imputation for covariates

  

Multiple imputation for covariates and outcome

  
  

(n = 1,112)

(n = 2,116)

(n = 4,129)

  

Estimation

se

p

Estimation

se

p

Estimation

se

p

Female gender

 

0.159

0.0282

<0.0001

0.134

0.0215

<0.0001

0.131

0.022

<0.0001

Age (/10 years)

 

0.044

0.020

0.027

0.023

0.0144

0.12

0.023

0.015

0.12

Intervention group

 

−0.028

0.022

0.20

−0.0146

0.0158

0.35

−0.015

0.017

0.40

Smoking status at inclusion

Never smoked

ref

 

0.13

ref

 

0.27

ref

 

0.18

Former smoker

0.029

0.0241

 

0.025

0.0175

 

0.0261

0.017

 

Smoker

0.069

0.035

 

0.032

0.027

 

0.0369

0.026

 

Number of dietary records

 

0.0013

0.0054

0.81

0.0007

0.0039

0.87

0.0007

0.0037

0.83

Energy intake (/1000, kcal/d)

 

−0.021

0.024

0.38

−0.04

0.017

0.023

−0.040

0.018

0.025

Alcohol intake at inclusion (log) (g/d)

 

0.0241

0.0082

0.0031

0.026

0.0061

<0.0001

0.023

0.0058

<0.0001

Plasma α-tocophérol (log) (μmol/L)

 

0.149

0.0414

0.00033

0.172

0.0358

<0.0001

0.177

0.0346

<0.0001

Plasma β-carotene (log) (μmol/L)

    

−0.034

0.0153

0.025

−0.0341

0.0149

0.024

Plasma Selenium (μmol/L)

 

0.1320

0.0598

0.027

0.103

0.045

0.021

0.106

0.041

0.010

Plasma Zinc (μmol/L)

 

0.0151

0.0055

0.0058

0.014

0.0043

0.0016

0.0129

0.0044

0.004

Catechins intake (mg/d)

> Q3

-

-

-

0.0798

0.0272

0.0033

0.034

0.017

0.047

Hydroxybenzoic acids intake (mg/d)

> Q3

   

−0.059

0.0268

0.0272

-

  
  1. aThe multivariable models first included eight subclasses of polyphenols (Dihydroflavonols, Proanthocyanidins, Theflavins, Catechins, Flavones, Flavonols, Hydroxybenzoic, Stilbenes) age, gender, intervention group, smoking status at inclusion, height and weight at inclusion, number of dietary records, energy intake, vitamin C, selenium, β-carotene, zinc, α-tocopherol, retinol, and alcohol intake at inclusion
  2. Bold data indicate statistical significance