Table 4.

Predictive characteristics of different cut-offs of lactate, pH and base excess for the composite neonatal morbidity (n=4910)

	Sensitivity (95% CI)	Specificity (95% CI)
Lactate (mmol/L)
>3.9 (‘optimal’ cut-off from ROC curve)	83.9 (71.7, 92.4)	74.1 (72.9, 75.4)
>4.8 [27]	62.5(48.5, 75.1)	87.6 (86.6, 88.5)
>6.0 [7]	46.4 (33.0, 60.3	94.9(94.2, 95.5)
>6.3 (>95th percentile)	0.0 (0.0, 6.4)	99.9 (99.8, 100)
pH
<7.00[28]	8.9(3.0, 19.6)	99.8 (99.6, 99.9)
<7.10 [29]	28.6 (17.3, 42.2)	98.9(98.6, 99.2)
<7.17 (<5th percentile)	42.9 (29.7, 58.8)	95.5 (94.9, 96.1)
<7.25 (‘optimal’ cut-off from ROC curve)	75.0 (61.6, 85.6)	70.6 (69.3, 71.9)
Base excess
> 4.3 (‘optimal’ cut-off from ROC curve)	71.4(57.8, 82.7)	79.5 (78.3, 80.6)
> 7.2 (>95th percentile)	46.4(33.0, 60.3)	95.7(95.1, 96.3)
> 12.0 [30]	19.6(10.2, 32.4)	99.6(99.4, 99.8)
pH<7.0 and base excess >12 [1]	8.9(3.0, 19.6)	99.8(99.6, 99.9)