Table 6

Sensitivity analysis of current best practice (stabilisation with NCPAP and high levels of antenatal steroid use)

30% Threshold40% Threshold50% Threshold60% Threshold
Mortality
Prophylaxis:
103 per 1000
(10.3%)		OR 1.03
(0.45 to 2.35)
Network estimate		2 more per 1000
(54 fewer to 110 more)		OR 1.32
(0.69 to 2.61)
Network estimate		29 more per 1000
(29 fewer to 127 more)		OR 0.81
(0.61 to 1.07)
Network estimate		18 fewer per 1000
(38 fewer to seven more)		OR 0.56
(0.23 to 1.29)
Network estimate		43 fewer per 1000
(78 fewer to 26 more)
Quality of evidence:
⊕◯◯◯ Very low *†‡§		Based on 279 participants
(1 RCT)		Based on 367 participants
(2 RCTs)		Based on 1469 participants
(2 RCTs)		Based on 439 participants
(1 RCT)
Bronchopulmonary dysplasia
Prophylaxis:
175 per 1000
(17.5%)		OR 1.40
(0.88 to 2.24)
Network estimate		54 more per 1000
(18 fewer to 148 more)		OR 0.83
(0.39 to 1.70)
Network estimate		26 fewer per 1000
(99 fewer to 91 more)		OR 0.93
(0.74 to 1.16)
Network estimate		11 fewer per 1000
(39 fewer to 22 more)		OR 1.29
(0.84 to 2.02)
Network estimate		41 more per 1000
(25 fewer to 125 more)
Quality of evidence:
⊕◯◯◯ Very low*†‡		Based on 279 participants
(1 RCT)		Based on 367 participants
(2 RCT)		Based on 1469 participants
(2 RCT)		Based on 439 participants
(1 RCT)
Pneumothorax
Prophylaxis:
27 per 1000
(2.7%)		OR 4.99
(0.00 to 6953.50)
Network estimate		94 more per 1000
(27 fewer to 968 more)		OR 3.09
(0.02 to 2455.29)
Network estimate		52 more per 1000
(26 fewer to 959 more)		OR 1.52
(0.01 to 324.08)
Network estimate		14 more per 1000
(27 fewer to 873 more)		OR 1.73
(0.00 to 2151.67)
Network estimate		19 more per 1000
(27 fewer to 957 more)
Quality of evidence:
⊕◯◯◯ Very low*†‡§		Based on 279 participants
(1 RCT)		Based on 367 participants
(2 RCTs)		Based on 1469 participants
(2 RCTs)		Based on 439 participants
(1 RCT)
Major morbidity
Prophylaxis:
296 per 1000
(29.6 per 100 participants)		RaR 1.20
(0.86 to 1.68)
Network estimate		60 more per 1000
(41 fewer to 202 more)		RaR 1.16
(0.81 to 1.66)
Network estimate		47 more per 1000
(56 fewer to 196 more)		RaR 1.06
(0.93 to 1.21)
Network estimate		19 more per 1000
(20 fewer to 62 more)		RaR 2.05
(1.46 to 2.93)
Network estimate		310 more per 1000
(136 more to 572 more)
Quality of evidence:
⊕◯◯◯ Very low*†‡§		Based on 279 participants
(1 RCT)		Based on 367 participants
(2 RCTs)		Based on 1469 participants
(2 RCTs)		Based on 439 participants
(1 RCT)
Grade 3/4 intraventricular haemorrhage
Prophylaxis:
39 per 1000
(3.9%)		OR 1.64
(0.24 to 14.41)
Network estimate		23 more per 1000
(29 fewer to 329 more)		OR 2.16
(0.87 to 5.98)
Network estimate		41 more per 1000
(5 fewer to 156 more)
Quality of evidence:
⊕◯◯◯ Very low*†‡§		Based on 279 participants
(1 RCT)		Based on 367 participants
(2 RCTs)		Based on 1469 participants
(2 RCTs)		Based on 439 participants
(1 RCT)

  • *The trials were all had some concerns or were at high risk of bias.

  • †There was significant heterogeneity.

  • ‡This is a surrogate outcome or was an indirect comparison.

  • §Less than 300 events in combined groups.

  • ¶There is evidence of publication bias.

  • NCPAP, nasal continuous positive airway pressure; RaR, rate ratio; RCT, randomised conrolled trial.