Secondary outcomes
Fourteen studies (8, 13-19, 22, 24, 26, 28, 36) that enrolled 2,083
neonates reported the duration of non-invasive ventilation, and found a
significant decrease in the duration of non-invasive ventilation using
NHFOV (standard mean difference (SMD) = -0.98, 95%CI -1.52, -0.45, I2 =
96%, p = 0.0003; suppl. eFig. 4). The subgroup analysis showed a
significant difference in post-extubation respiratory support in the
cohort studies and RCTs (SMD = -0.47, 95%CI -0.84, -0.09, I2 = 0%, p =
0.01; SMD = -1.52, 95%CI -2.58.-0.45, I2 = 98%, p = 0.005), whereas no
significant differences were observed in initial respiratory support
(SMD = -0.15, 95%CI -0.49, 0.19, I2 = 41%, p = 0.40) (SMD = -1.20,
95%CI -2.88, 0.48, I2 = 97%, p = 0.16).
Ten studies (9, 14, 16-18, 24, 26, 28, 36) with 1,792 neonates that
reported the total oxygen therapy time found no significant differences
between the NHFOV and NIPPV groups (SMD = -0.08, 95%CI -0.28, 0.12, I2
= 69%, p = 0.45; suppl. eFig. 5). No significant differences were found
in the subgroup analysis of two cohort studies and two RCTs in the total
oxygen therapy time in the initial respiratory support (SMD = -0.23,
95%CI -0.71, 0.25, I2 = 67%, p = 0.34) (SMD = -0.11, 95%CI -0.42,
0.20, I2 = 0%, p = 0.47). Nor were significant differences found among
the pooled data of five RCTs post-extubation respiratory support between
NHFOV and NIPPV (SMD = -0.01, 95%CI -0.37, 0.35, I2 = 84%, p = 0.95).
Nine studies (7, 9, 14, 16, 18, 24, 26, 27) with 948 neonates that
reported LOS showed a significant difference in the decreased LOS (SMD =
-0.24, 95%CI -0.47, -0.01, I2 = 66%, p =
0.04; suppl. eFig. 6). The subgroup analysis showed no significant
difference in the initial respiratory support of the pooled data between
the cohort study and RCTs (SMD = -0.26, 95%CI -0.72, 0.19,
I2 = 84%, p = 0.26; SMD = -0.06, 95%CI -0.37,
0.25, I2 = 0%, p = 0.69).
Three trials (24, 25, 27) with 291 neonates reported the results of
blood gas analyses (PaO2 and PaCO2 levels and SpO2/FiO2 ratios) 1 h
after initial non-invasive respiratory support. The NHFOV significantly
reduced PaCO2 levels (SMD = -1.48, 95%CI -1.74, -1.22,
I2 = 0%, p <0.001) and increased PaO2 levels
(SMD = 0.42, 95%CI 0.19, 0.65, I2 = 0%, p< 0.001) and the SpO2/FiO2 ratio (SMD = 0.47, 95%CI 0.24,
0.70, I2 = 0%, p < 0.001) in
neonates, unlike NIPPV (suppl. eTable 1).
Seven trials (14, 15, 17, 18, 20, 29) enrolling 540 neonates reported
PaCO2 levels 24 h after non-invasive respiratory support. According to
the meta-analysis, NHFOV (but not NIPPY) reduced PaCO2 levels
significantly (SMD = -0.64, 95%CI -0.92, -0.36, I2 =
60%, p <0.001) in neonates. Six trials (14, 15, 17, 18, 29)
reported PaO2 levels and the meta-analysis indicated that NHFOV
significantly enhanced PaO2 levels (SMD = 0.40, 95%CI 0.14, 0.67,
I2 = 55%, p = 0.003) compared with NIPPV.
Significant differences were found only in the pooled data of RCTs of
initial respiratory support (SMD = 0.57, 95%CI 0.26, 0.88,
I2 = 45%, p < 0.001. Five trials
(14, 15, 18, 29) that reported SpO2/FiO2 ratios, found that NHFOV
significantly enhanced the SpO2/FiO2 ratio (SMD = 0.56, 95%CI 0.29,
0.83, I2 = 29%, p <0.001). The subgroup
analysis showed a significant difference only in the pooled RCT data for
initial respiratory support (SMD = 0.42, 95%CI 0.07, 0.76,
I2 = 66%, p = 0.02;suppl. eTable 1).
Seventeen studies (13-17, 19, 22-24, 26-29, 36) with 2,491 neonates
reported adverse outcomes, including the incidence of BPD, and showed
that NHFOV reduced the risk of BPD (OR = 0.77, 95%CI 0.63, 0.94, I2 =
0%, p = 0.01; suppl. eTable 2). In the subgroup analysis, significant
differences were observed in the meta-analyses of RCTs for initial
respiratory support (OR = 0.49, 95% CI 0.25, 0.94, I2 = %, p = 0.03).
Seven studies (14, 17, 23, 24, 27, 28) with 674 neonates reported the
incidence of apnea, and showed that the NHFOV resulted in a significant
reduction in apnea (OR = 0.55, 95%CI 0.34, 0.88, I2 =
0%, p = 0.01;suppl. eTable 2). However, no significant
differences were found between NHFOV and NIPPV after a subgroup analysis
by study type and initial or post-extubation respiratory support.
Furthermore, NHFOV reduced the incidence of abdominal distention in
post-extubation respiratory support in cohort studies (OR = 0.22, 95%CI
0.07, 0.71, I2 = 0%, p = 0.01; suppl. eTable
2). No significant differences in the likelihood of other adverse
outcomes (including air leaks, NEC, IVH, nasal injury, ROP, PNX and
periventricular leukomalacia) were observed ( suppl. eTable 2).