Table 2.
Distribution of baseline characteristics propensity score in different cohort subpopulations.
| Groups and Values | p | ||
|---|---|---|---|
| PIRA | No PIRA | ||
|
Propensity Score (SD) |
0.519 (0.058) |
0.480 (0.082) |
0.075 |
| Disease duration lowest quartile (≤ 2.6 years) | Disease duration highest quartile (≥ 15.3 years years) | ||
|
Propensity Score (SD) |
0.527 (0.053) |
0.471 (0.079) |
0.038 |
| Number of DMTs prior to OCR lower than median (2) | Number of DMTs prior to OCR higher than median (2) | ||
|
Propensity Score (SD) |
0.499 (0.071) |
0.476 (0.087) |
0.310 |
| First-line OCR | Second-line OCR | ||
|
Propensity Score (SD) |
0.480 (0.084) |
0.505 (0.066) |
0.301 |
| CD19+ Not Detectable | CD19+ Detectable | ||
|
Propensity Score (SD) |
0.490 (0.080) |
0.472 (0.079) |
0.376 |
| Smallest pRNFL lowest quartile (≤ 71 µm) | Smallest pRNFL highest quartile (≥ 91 µm) | ||
|
Propensity Score (SD) |
0.479 (0.106) |
0.502 (0.078) |
0.108 |
| Smallest mRNFL lowest quartile(≤ 0.69 mm3) | Smallest mRNFL highest quartile (≥ 0.91 mm3) | ||
|
Propensity Score (SD) |
0.453 (0.092) |
0.504 (0.056) |
0.142 |
| Smallest GCIPL lowest quartile (≤ 1.53 mm3) | Smallest GCIPL highest quartile (≥ 1.82 mm3) | ||
|
Propensity Score (SD) |
0.465 (0.093) |
0.510 (0.077) |
0.158 |
A propensity score for the development of PIRA was calculated using logistic regression taking into account the baseline characteristics sex, EDSS at baseline, ARR at baseline and age at ocrelizumab treatment. A higher score indicates a higher likelihood for baseline characteristics that favor PIRA development. The table compares the scores of different subgroups. Significance of differences between subgroups was tested using a Mann–Whitney-U test. Differences were considered statistically significant with the following p-values: *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001.