Table 3 Predicted rates of annual retinal thinning and brain volume loss in relation to inflammatory activity and time from MS onset.

From: Retinal and brain damage during multiple sclerosis course: inflammatory activity is a key factor in the first 5 years

Years from MS onset

Area

Activity

No activity

Ratio activity/no activity (95% CI)

Year 1

GCIPL

2.24

1.21

1.84 (1.16; 3.11)

pRNFL

1.26

1.62

1.28 (0.72; 2.14)

Whole brain

1.32

0.57

2.32 (1.48; 4.22)

Gray matter

1.10

0.67

1.64 (0.93; 2.98)

Thalamus

2.56

0.47

5.29 (2.06; 32.71)

Year 2

GCIPL

1.81

1.03

1.74 (1.15; 2.75)

pRNFL

1.16

1.42

1.22 (0.76; 1.84)

Whole brain

0.96

0.52

1.84 (1.31; 2.68)

Gray matter

0.84

0.58

1.44 (0.96; 2.17)

Thalamus

1.92

0.59

3.27 (2.08; 6.58)

Year 3

GCIPL

1.37

0.85

1.61 (1.13; 2.37)

pRNFL

1.06

1.22

1.15 (0.81; 1.59)

Whole brain

0.70

0.49

1.43 (1.10; 1.89)

Gray matter

0.65

0.53

1.25 (0.91; 1.69)

Thalamus

1.46

0.68

2.15 (1.59; 3.14)

Year 4

GCIPL

0.67

0.93

1.39 (1.01; 1.98)

pRNFL

0.96

1.02

1.07 (0.83; 1.36)

Whole brain

0.52

0.47

1.13 (0.86; 1.46)

Gray matter

0.53

0.49

1.09 (0.81; 1.48)

Thalamus

1.14

0.74

1.53 (1.18; 2.09)

Year 5

GCIPL

0.49

0.50

1.01 (0.52; 1.88)

pRNFL

0.85

0.82

0.96 (0.70; 1.29)

Whole brain

0.42

0.45

0.94 (0.69; 1.24)

Gray matter

0.47

0.46

0.98 (0.69; 1.35)

Thalamus

0.94

0.79

1.2 (0.90; 1.61)

Year 10

GCIPL

0.54

0.58

1.09 (0.77; 1.52)

pRNFL

0.79

0.70

0.99 (0.80; 1.22)

Whole brain

0.44

0.43

1.04 (0.80; 1.32)

Gray matter

0.48

0.48

1.01 (0.77; 1.31)

Thalamus

0.84

0.77

1.14 (0.89; 1.46)

Year 15

GCIPL

0.59

0.583

1.15 (0.84; 1.59)

pRNFL

0.71

0.73

1.03 (0.79; 1.32)

Whole brain

0.42

0.32

1.33 (0.97; 1.86)

Gray matter

0.47

0.39

1.20 (0.87; 1.65)

Thalamus

0.88

0.58

1.51 (0.13; 2.15)

  1. The data represents the predicted annualized rates (%/year) of retinal thinning and brain volume loss using linear splines mixed effect models. In addition to estimates, we added ratios to ease interpretation. The 95% confident intervals were calculated by parametric bootstraps and significant differences are indicated in bold.