Skip to main content

Table 2 Logistic regression of B12 deficiency on potential determinants (n = 592)

From: Determinants of vitamin B12 deficiency in patients with type-2 diabetes mellitus — A primary-care retrospective cohort study

Variables

Univariate logistic regression

Multivariable logistic regression

OR (95% CI)

P value

OR (95% CI)

P value

Age groupa (Reference: < 60 years)

 

0.259 +

 

0.027

60–79 years

0.70 (0.46 – 1.08)

0.105

1.26 (0.77 – 2.07)

0.357

≥ 80 years

0.84 (0.44 – 1.61)

0.606

2.86 (1.31 – 6.25)

0.008

Vegetariana (Reference = non-vegetarian)

16.82 (3.72 – 76.00)

< 0.001

21.61 (4.47 – 104.44)

< 0.001

B12 supplementation (prescribed)a

0.40 (0.26 – 0.62)

< 0.001

0.37 (0.22 – 0.61)

< 0.001

Metformin daily dosea (g/day)

2.44 (2.00 – 2.97)

< 0.001

2.79 (2.22 – 3.49)

< 0.001

Folate deficiencya

2.02 (1.34 – 3.04)

< 0.001

2.04 (1.27 – 3.28)

0.003

LDL cholesterol levela (mmol/L)

0.73 (0.54 – 0.98)

0.031

  

Ethnicitya (Reference: Chinese)

 

0.220 +

  

Malay

0.85 (0.41 – 1.78)

0.668

  

Indian

1.63 (0.97 – 2.75)

0.065

  

Others

0.72 (0.26 – 1.96)

0.516

  

Male gendera (Reference = Female)

1.19 (0.83 – 1.70)

0.353

  

Duration of diabetes for ≥ 10 yearsa

1.55 (1.08 – 2.24)

0.017

  

Calcium supplementation (prescribed)a

0.52 (0.35 – 0.77)

< 0.001

  

PPI or H2A usea

0.93 (0.60 – 1.46)

0.760

  

Sulphonylurea usea

2.58 (1.77 – 3.76)

< 0.001

  

Acarbose usea

3.46 (2.16 – 5.55)

< 0.001

  

HbA1ca (%)

1.26 (1.07 – 1.48)

< 0.01

  

Statin use

1.20 (0.72 – 2.00)

0.481

  

Fenofibrate use

1.46 (0.83 – 2.57)

0.192

  

Insulin use

1.60 (0.90 – 2.83

0.114

  

DPP4-inhibitor use

0.82 (0.34 – 1.97)

0.656

  

Creatinine (µmol/L)

1.00 (0.99 – 1.01)

0.966

  

eGFR (mL/min/1.73m2)

1.00 (0.99 – 1.01)

0.694

  

Hypertension

0.89 (0.52 – 1.50)

0.651

  

Hyperlipidaemia

1.16 (0.60 – 2.24)

0.645

  

Body Mass Index (kg/m2)

1.02 (0.98 – 1.06)

0.328

  
  1. Abbreviation: eGFR, estimated glomerular filtration rate adjusted for body surface area
  2. aPredictor variables that were further analysed within multivariable logistic regression. + refers to omnibus p – value. “Age”, “Ethnicity”, “Gender” and “PPI or H2A use” were predetermined to be clinically or demographically relevant, and thus forced into the modelling process