Logistic regression analyses for outcome of being CMV positive (CMV+) in marrow and for outcome of possessing gB3 among patients who were CMV+ in marrow
| . | Odds ratio . | 95% CI . | P . |
|---|---|---|---|
| Group (CMV+/total) | |||
| Controls (19/151) | 1 | — | — |
| AA (33/100) | 3.45 | 1.70-7.00 | .0006 |
| Group (gB3+/total CMV+) | |||
| Controls (5/19) | 1 | — | — |
| AA (19/33) | 7.83 | 1.72-35.79 | .008 |
| . | Odds ratio . | 95% CI . | P . |
|---|---|---|---|
| Group (CMV+/total) | |||
| Controls (19/151) | 1 | — | — |
| AA (33/100) | 3.45 | 1.70-7.00 | .0006 |
| Group (gB3+/total CMV+) | |||
| Controls (5/19) | 1 | — | — |
| AA (19/33) | 7.83 | 1.72-35.79 | .008 |
Models were adjusted for age. To compensate for a potential overrepresentation of stroma in AA biopsies compared with control aspirates, the aspirated marrow was also cultured to expand the stromal cells. Both aspirated marrow cells and expanded stromal cells were then evaluated by PCR. Combining both tests did not change the frequency of CMV+ samples in the control group.