Table 4. Complications of Prematurity and Other Serious Adverse Events Documented After Randomization.
| Events | No./total (%) | Absolute difference, % (95% CI) | |
|---|---|---|---|
| Liberal transfusion threshold (n = 492) | Restrictive transfusion threshold (n = 521) | ||
| Any intraventricular/periventricular hemorrhagea | 117/492 (23.8) | 113/521 (21.7) | 2.1 (−3.1 to 7.3) |
| Intraventricular/periventricular hemorrhage grade 3 or 4a | 40/492 (8.1) | 35/521 (6.7) | 1.4 (−1.8 to 4.6) |
| Cystic periventricular leukomalaciab | 23/492 (4.7) | 30/521 (5.8) | −1.1 (−3.8 to 1.7) |
| Bronchopulmonary dysplasiac | 130/458 (28.4) | 126/485 (26.0) | 2.4 (−3.3 to 8.1) |
| Necrotizing enterocolitisd | 26/492 (5.3) | 32/518 (6.2) | −0.9 (−3.8 to 2.0) |
| Surgical necrotizing enterocolitis | 20/492 (4.1) | 28/518 (5.4) | −1.3 (−4.0 to 1.3) |
| Focal intestinal perforation | 31/492 (6.3) | 33/518 (6.4) | −0.1 (−3.1 to 2.9) |
| Retinopathy of prematurity | |||
| Any | 257/472 (54.5) | 261/492 (53.1) | 1.4 (−4.9 to 7.7) |
| Maximum stage ≥3 | 75/472 (15.9) | 64/492 (13.0) | 2.9 (−1.6 to 7.3) |
| Receiving therapy | 41/472 (8.7) | 38/492 (7.7) | 1.0 (−2.5 to 4.4) |
| Patent ductus arteriosus | |||
| Any therapy | 204/492 (41.5) | 196/518 (37.8) | 3.6 (−2.4 to 9.7) |
| Ligation/any therapy | 51/204 (25.0) | 52/196 (26.5) | −1.5 (−10.1 to 7.0) |
| Culture-proven sepsis excluding coagulase-negative staphylococcie | |||
| Any | 59/492 (12.0) | 61/518 (11.8) | 0.2 (−3.8 to 4.2) |
| No. of episodes | |||
| 0 | 433 (88.0) | 457 (88.2) | |
| 1 | 50 (10.1) | 51 (9.8) | |
| 2 | 6 (1.2) | 8 (1.5) | |
| 3 | 3 (0.6) | 2 (0.4) | |
| Culture-proven sepsis with coagulase-negative staphylococcie | |||
| Any | 54/492 (11.0) | 62/518 (12.0) | −1.0 (−4.9 to 2.9) |
| No. of episodes | |||
| 0 | 438 (89.0) | 456 (88.0) | |
| 1 | 48 (9.8) | 57 (11.0) | |
| 2 | 5 (1.0) | 5 (1.0) | |
| >2 | 1 (0.2) | 0 | |
| Clinical sepsise | |||
| Any | 142/492 (28.9) | 153/518 (29.5) | −0.7 (−6.3 to 4.9) |
| No. of clinical sepsis episodes | |||
| 0 | 350 (71.1) | 365 (70.5) | |
| 1 | 98 (19.9) | 112 (21.6) | |
| 2 | 32 (6.5) | 27 (5.2) | |
| 3 | 9 (1.8) | 9 (1.7) | |
| >3 | 3 (0.6) | 5 (1.0) | |
| Pneumoniae | |||
| Any | 30/492 (6.1) | 32/518 (6.2) | −0.1 (−3.0 to 2.9) |
| No. of pneumonia episodes | |||
| 0 | 462 (93.9) | 486 (93.8) | |
| 1 | 26 (5.3) | 23 (4.4) | |
| 2 | 3 (0.6) | 4 (0.8) | |
| 3 | 1 (0.2) | 2 (0.4) | |
| >3 | 0 | 3 (0.6) | |
| No. of additional serious adverse eventsf | 15 | 29 | |
| No. of potentially related serious adverse eventsg | 3 | 1 | |
Data on intraventricular/periventricular hemorrhage include findings already evolving before randomization. Findings of prerandomization head ultrasound are shown in eTable 5 in the Supplement.
Data on cystic periventricular leukomalacia include findings already evolving before randomization. Rates of cystic periventricular leukomalacia at baseline were 1 (0.2%) and 5 (1%) in the liberal and restrictive threshold groups, respectively.
Bronchopulmonary dysplasia was determined based on need for positive pressure respiratory support or supplemental oxygen at 36 weeks of postmenstrual age, including a room air test if indicated. The rates of death before 36 weeks of postmenstrual age precluding the diagnosis by definition were 29 (6%) and 30 (6%) in the liberal and restrictive threshold groups, respectively.
Necrotizing enterocolitis was diagnosed if criteria for modified Bell stage ≥IIa were present.
Blood culture–proven sepsis, clinical sepsis, and pneumonia were recorded according to the German Nosocomial Infection Surveillance System for Preterm Infants in Neonatology Departments and ICU Infection Surveillance (Neo-KISS) criteria (https://www.nrz-hygiene.de/en/surveillance/hospital-infection-surveillance-system/neo-kiss/), with data given for the number of separate episodes of illness meeting the diagnostic criteria.
Additional serious adverse events limited to those reported in >2 infants included, in the liberal vs restrictive threshold groups, respectively: volvulus (n = 3 vs n = 4), cytomegalovirus infection (n = 1 vs n = 4), meningitis (n = 2 vs n = 4), and renal failure (n = 1 vs n = 2).
Four serious adverse events (1 necrotizing enterocolitis, 1 cytomegalovirus infection, and 2 sepsis) were considered by local investigators to be possibly or probably related to the assigned threshold or a preceding red blood cell transfusion (neither the cytomegalovirus infection nor the sepsis were microbiologically proven to be related to the red blood cell transfusion).