Intrapartum fetal stimulation tests: a meta-analysis

To assess the performance of stimulation tests for the prediction of intrapartum fetal acidemia.We conducted a MEDLINE (Internet Grateful Med) literature review from 1966 to 2000 using the terms "fetal scalp pH," "fetal scalp stimulation," and "fetal acoustic stimulation."Articles were included if sensitivity, specificity, and predictive values for intrapartum fetal acidemia could be calculated. Reactivity was a fetal heart rate (FHR) acceleration of 15 beats per minute for 15 seconds. Likelihood ratio and 95% confidence intervals (CIs) for four different fetal provocations were calculated using the Cochrane collaboration 2000 Review Manager 4.1. This permitted an estimate of the degree of confidence surrounding the point estimate of the likelihood ratio for the presence or absence of acidemia given a positive or negative test. The likelihood ratio is a stable predictive property of any test because it combines information from both sensitivity and specificity, is independent of prevalence, and avoids the limitations of traditional predictive values.Eleven of 512 articles met criteria for inclusion and included four stimulation tests - fetal scalp puncture, Allis clamp scalp stimulation, vibroacoustic stimulation, and digital scalp stimulation. Pooled likelihood ratio and 95% CIs were similar among the four different stimulation tests. Each test was very useful at predicting both the lack of and the presence of fetal acidemia. Likelihood ratio and 95% CIs for the prediction of fetal acidemia given a positive test were: scalp puncture 8.54 (CI 1.28, 56.96), Allis clamp 10.4 (CI 1.47, 73.61), vibroacoustic stimulation 5.06 (CI 2.69, 9.50), and digital 15.68 (CI 3.22, 76.24). For a negative test, these were: scalp puncture 0.12 (CI 0.02, 0.78), Allis clamp 0.10 (CI 0.01, 0.68), vibroacoustic stimulation 0.20 (CI 0.11, 0.37), and digital 0.06 (CI 0.01, 0.31).Intrapartum stimulation tests appear to be useful to rule out fetal acidemia in the setting of a nonreassuring FHR pattern. Our data reveal the degree of confidence around the estimate of the likelihood ratio of a stimulation test. The very low negative likelihood ratios warrant the use of these tests when a nonreassuring intrapartum FHR pattern appears. Because these tests are less than perfect, caution is advised; careful continued monitoring with repeat testing during the course of labor should be performed as long as suspicious FHR patterns persist. Fetal scalp pH should be determined whenever possible after a positive stimulation test (lack of acceleration).