Background: [corrected] Cardiotocography (CTG) seems to be a non-reliable, expensive but nevertheless practical method for fetal surveillance. Moreover, its diagnostic value is dependent on a long-standing experience of the obstetrician (midwife). It is difficult to define exact diagnostic criteria since nearly all CTG phenomena are lacking precise qualification by the naked eye. Therefore, the idea must be born in mind to analyze fetal heart frequency (FHF) by computer, first off-line then on-line, in order to evaluate its true diagnostic power objectively.
Methods: The FHF of 583 deliveries terminated by the vaginal route were registered prospectively using a PC and an RS422 interface. In 443 cases acid-base measurements (ABL 500, RADIOMETER, Copenhagen) in blood of the umbilical artery (UA) and vein (UV) were available and plausible. In this study only the last 30 min ante-partum were analyzed. The program for FHF analysis was written in MATLAB (The Mathworks Inc., USA). A CTG score was developed using three components of FHF: basal FHF, the deceleration area of all dips, and the micro-fluctuation (MF) of the basal fetal heart rate (FHR). MF denotes the true number of "turning-points" per minute of basal FHR. For each component a maximum of 6 scoring points could be assigned according to empirical cut-off values. These cut-off values were determined using correlation analysis with acid-base parameters in UA blood, especially the actual pH. The accordance between score and pH values was further demonstrated by assignment of 0.036 pH-units to each of the 19 (18 + zero) scoring points, thus covering a pH range between 6.700 and 7.350 (UA). A resulting variable, delta pH (pH measured, UA-pH assigned) was studied and used for further analysis. In order to define criteria for fetal mortality in utero only cases with pH, UA between 7.250 and 7.350 were accepted.
Results: Median basal FHF under normal conditions amounted to 138 bpm (mean: 137 +/- 14.9) in 4180 minutes of 372 fetuses. 120 bpm equals the 13.4(th) and 160 bpm the 94.6(th) centile of the distribution. Given fetal normacidity (UA) MF is 58/minute and the mean MF 57.9 +/- 13.4, respectively, with a 10 (th) centile of 41/minute. MF and basal FHF are correlated significantly (r = 0.410, P << 10(-4)). The declaration-area per fetus is significantly correlated with actual pH (UA), r = -0.473, P << 10(-4). the score itself is highly significantly correlated with actual pH (UA) (r = -0.559, P << 10(-4)) and the other parameters of fetal acid-base balance. Nevertheless, prediction variability for pH, especially in score = 1, 2 and zero (minimal CTG pathology) is still present: 80% of all predicted pH values lie in between -0.092 and + 0.071. It is strongly suggested that this score-related predictive pH variability is caused by maternal breathing habits during the last 30 minutes of delivery.
Conclusion: Adequate quantification of only three variables of FHR using a score leads to fairly good correlations with parameters of the fetal acid-base balance. Thus actual pH (UA) can be predicted in reasonable clinical limits. Still present variability in prediction of pH seems to be, in part, of maternal origin. The maternal influence could be eliminated by continuous (transcutaneous) monitoring of maternal pCO(2). Along these lines the quantitative electronic monitoring of FHR will be realized and instrumented (off-line and on-line) by nexus/gmt, Frankfurt, a.M., Germany.