Table 3.
Predefined secondary outcomes during neonatal hospitalisation
| Hypothermia (n=202) | Control (n=206) | Risk ratio*(95% CI) | p value | |
|---|---|---|---|---|
| Intracranial haemorrhage | 2 (1%) | 4 (2%) | 0·51 (0·09–2·75) | 0·68 |
| Gastric bleeding | 62 (31%) | 34 (17%) | 1·86 (1·28–2·69) | <0·00070 |
| Persistent hypotension | 45 (22%) | 25 (12%) | 1·84 (1·17–2·88) | 0·0066 |
| Pulmonary haemorrhage | 42 (21%) | 28 (14%) | 1·53 (0·99–2·37) | 0·054 |
| Persistent pulmonary hypertension | 24 (12%) | 16 (8%) | 1·53 (0·84–2·79) | 0·16 |
| Prolonged blood coagulation | 79 (39%) | 52 (25%) | 1·55 (1·16–2·07) | 0·0027 |
| Culture-positive early-onset sepsis† | 12 (6%) | 10 (5%) | 1·22 (0·54–2·77) | 0·63 |
| Necrotising enterocolitis | 5 (2%) | 1 (0%) | 5·10 (0·60–43·2) | 0·12 |
| Cardiac arrhythmia | 5 (2%) | 0 | .. | 0·029 |
| Severe thrombocytopenia | 33 (16%) | 15 (7%) | 2·24 (1·26–4·00) | 0·0045 |
| Persistent metabolic acidosis | 46 (23%) | 24 (12%) | 1·95 (1·24–3·08) | 0·0029 |
| Renal failure | 22 (11%) | 16 (8%) | 1·40 (0·76–2·59) | 0·28 |
| Pneumonia | 26 (13%) | 25 (12%) | 1·06 (0·63–1·77) | 0·82 |
| Subcutaneous fat necrosis | 1 (0%) | 0 | .. | .. |
| Hospital stay,‡ days, median (IQR) | 16·1 (12·9–23·2) | 13·9 (11·0–18·8) | 2·20 (0·70–3·80)¶ | 0·0044 |
| Death before discharge | 72 (36%) | 49 (24%) | 1·50 (1·10–2·04) | 0·0087 |
| Neurological examination at discharge§ | 50 (39%) | 65 (41%) | 0·93 (0·70–1·24) | 0·61 |
Data are number of patients (%), unless otherwise indicated. There were no missing data. Intracranial haemorrhage refers to a major parenchymal or intraventricular bleed on cranial ultrasound. Gastric bleeding refers to more than 5 mL fresh blood produced from the nasogastric tube. Persistent hypotension refers to a mean blood pressure of less than 25 mm Hg, despite maximum inotropic support. Pulmonary haemorrhage refers to copious bloody secretions with clinical deterioration requiring change(s) in ventilatory management. Persistent pulmonary hypertension refers to severe hypoxaemia disproportionate to the severity of lung disease with a significant pre-ductal and post-ductal saturation difference on pulse oximetry. Long-term blood coagulation refers to abnormal coagulation times requiring the administration of blood products. Culture proven early-onset sepsis refers to the isolation and identification of a pathogenic organism from blood or cerebrospinal fluid, or both, along with clinical evidence of sepsis within 72 h of birth. Necrotising enterocolitis refers to abdominal distension, increased gastric aspirates, or blood in stools, or a combination, together with an abdominal x-ray showing bowel oedema, pneumatosis or pneumoperitoneum. Cardiac arrythmias included ventricular arrythmias and ectopic heartbeats requiring treatment. Severe thrombocytopenia refers to a platelet count of less than 25 000 per μL or less than 50 000 per μL with active bleeding. Persistent metabolic acidosis refers to a blood pH of less than 7·15 for more than 12 h with a normal partial pressure of carbon dioxide. Renal failure refers to anuria lasting longer than 48 h with elevated creatinine. Pneumonia refers to infiltrates on a chest x-ray consistent with infection or aspiration.
Group difference expressed as the risk ratio of occurrence in hypothermia group relative to the control group.
In the hypothermia group, seven infants had Klebsiella pneumoniae infections, two had Pseudomonas spp, one had Enterobacter spp, one had Escherichia coli, and one had non-fermenting Gram-negative bacilli isolated from blood culture; eight were born at the participating hospital and four were referred to the participating hospital. Among the infants in the control group, five had K pneumoniae infections, three had E coli, and two had non-fermenting Gram-negative bacilli isolated from blood culture; four were born at the participating hospital and six were referred to the participating hospital.
Figures based 130 hypothermia babies, 156 control babies who survived to discharge (1 missing value for surviving babies in control group, 0 missing in hypothermia group).
Figures based on babies surviving to discharge only. Figures based 130 hypothermia babies, 157 control babies.
Median difference (95% CI).