Table 5.
Indications for surgery
| Indications for surgery | ||||||
|---|---|---|---|---|---|---|
| Sometimes | Always | |||||
| n [%] | (95% CI) | p | n [%] | (95% CI) | p | |
| Pneumoperitoneum on AR, n = 202 | 22 [11%] | (7–16%) | 178 [88%] | (82–92%) | ||
| Pneumoperitoneum on US … all, n = 202 |
51 [25%] | (20–32%) | < 0.001e | 101 [50%] | (43–57%) | < 0.001e |
| … where US was used for NEC, n = 118 | 37 [31%] | (23–40%) | 72 [61%] | (52–69%) | ||
| … where US was not known to be used for NEC, n = 84 | 15 [18%] | (11–28%) | 29 [35%] | (25–45%) | 0.037f | |
| Portal venous gas on AR, n = 202 | 93 [46%] | 38 [19%] | ||||
| Portal venous gas on USa … all, n = 202 |
79 [39%] | 22 [11%] | ||||
| … where US was used for NEC, n = 118 | 54 [46%] | 15 [13%] | ||||
| … where US was not known to be used for NEC, n = 84 | 25 [30%] | 7 [8%] | ||||
| Fixed/persistent loop on sequential AR, n = 202b | 121 [60%] | 30 [15%] | ||||
| Turbid or localized fluid on USc … all, n = 202 |
114 [56%] | 24 [12%] | ||||
| … where US was used for NEC, n = 118 | 80 [68%] | (59–76%) | 18 [15%] | (10–23%) | ||
| … where US was not known to be used for NEC, n = 84 | 34 [40%] | (30–51%) | < 0.001f | 6 [7%] | (3–15%) | |
| Clinical deterioration despite medical treatment, n = 128d | 52 [41%] | 72 [56%] | ||||
Ninety-five per cent confidence intervals and p values are specified where there were significant differences between modalities or respondents with and without experience of US in NEC
a)A significantly greater proportion of surgeons than radiologists regarded portal venous gas on ultrasound as “sometimes” an indication for surgery: 52% (39-65%) vs. 26% (17-37%), p = 0.002. Summarising all positive responses (always and sometimes), the difference was significant between PVG on AR, 65% (58-72%), and on US 50% (43-57%), p = 0.003
b)A significantly greater proportion of radiologists than surgeons rejected persistent loop as an indication for surgery: 31% (21-43%) vs. 7% (3-17%), p = 0.001. Supplementary analyses showed no substantial influence of geographical variations
c)A significantly greater proportion of neonatologists than radiologists rejected turbid or localised fluid as an indication for surgery: 36% (26-48%) vs. 9% (4-19%), p < 0.001. There was no substantial influence of the geographical locations of the hospitals of the respondents
d)Included in the clinicians' questionnaire only
e)PP on AR compared to PP on US
f)Where ultrasound was known to be used in NEC compared to where it was not