Skip to main content
. 2021 Jul 7;36(7):1461–1467. doi: 10.1038/s41433-021-01672-1

Table 2.

Mixed effects linear regression model for 12-month visual acuity in DMO using central subfield.

Factors Estimates CI p value
(Intercept) 46.40 30.91 61.90 <0.001
CST SD across 12 months (100s µm) −6.87 −10.52 −3.23 <0.001
Baseline visual acuity (ETDRS letters) 0.35 0.27 0.44 <0.001
Baseline CST (100s µm) 1.63 0.18 3.09 0.027
Age (years) −0.03 −0.15 0.10 0.682
Gender: male (vs. female) 2.96 0.49 5.44 0.019
DR stage: PDR (vs. mild–moderate NPDR) −2.20 −5.63 1.22 0.207
DR stage: severe NPDR (vs. mild–moderate NPDR) −3.95 −7.35 −0.56 0.023
HbA1c (%) 0.24 −0.43 0.91 0.484
Laser: focala (vs. none) −1.85 −7.73 4.03 0.538
Laser: PRPb (vs. none) −1.29 −4.27 1.69 0.396
Laser: both focal and PRPc (vs. none) −4.45 −12.63 3.73 0.286
Number of anti-VEGF injections 0.38 −0.18 0.95 0.183
Insulin: dependent (vs. non-dependent) −0.13 −2.96 2.69 0.926
Anti-VEGF injection type: bevacizumab (vs. aflibercept) −5.33 −10.39 −0.26 0.039
Anti-VEGF injection type: mixedd (vs. aflibercept) −7.38 −12.76 −2.00 0.007
Anti-VEGF injection type: ranibizumab (vs. aflibercept) −4.02 −25.01 16.97 0.707
Model characteristics
σ2 65.86
 τ00 id 77.77
 ICC 0.54
 Marginal R2/conditional R2 0.273/0.667

Estimates denote the predicted change in visual acuity at 12 months for a 1 unit increase of a numeric factor, or for that factor level relative to reference level of a categorical factor, with other factors held constant.

CI confidence interval, CST central subfield thickness, SD standard deviation, σ2 variance within subjects, τ00 id variance between subjects, ICC intraclass correlation coefficient, R2 coefficient of determination.

aFocal laser refers to eyes that had received focal laser prior to the study period.

bPRP refers to eyes that had received PRP either prior to or during the study period.

cBoth focal and PRP refers to eyes that had received focal laser prior to the study period in addition to PRP at any time.

dMixed refers to eyes that had received a combination of bevacizumab and aflibercept.

Statistically significant p < 0.05 values are in bold.