Townsend, Barbis, and Mathews
operative procedures and even angiographic embolization, which in itself is not totally benign.
September 1991 Am J Obstet Gynecol
2. Druzin ML. Packing of lower uterine segment for control of postcesarean bleeding in instances of placenta previa. Surg Gynecol Obstet 1989;169:543-5.
REFERENCES 1. King PA, Duthie Sj, Dog ZG, Ma HK. Secondary postpartum haemorrhage. Aust NZ j Obstet Gynaecol 1989;29:394-8.
Computerized analysis of episodic changes in fetal heart rate variation in early labor G.S. Dawes, DM, S.K. Rosevear, MD, L.C. Pello, MD, M. Moulden, and C.W.G. Redman, MD Oxford, England Fetal heart rate variation in early labor was measured by computerized analysis in cyclic episodes of low or high variation in 136 women at 37 to 42 weeks' gestation. The amplitude (mean ± SE) in episodes of low variation was 20.6 ± 0.4 milliseconds; in high variation it was 57.3 ± 1.1 milliseconds. The duration (mean ± SE) of low episodes (24.3 ± 1.3 minutes) was less than that of high episodes (45.1 ± 2.7 minutes) but was sometimes >1 hour. In episodes of low variatibn the amplitude was 25% FHR 200
}-=!,V''''''';,J''~;,.J-i-''-..;j',-:.......j,'-:.~-:·.}.f\.,,,,J',,.··'\-7~·j::'',"",-.••i",~,·,t"-/.~-J.~j" "'~""""·.f:"'-""'----: o
--....------ - ....--
Fig. 1. Computer analysis of 2 hours in the early first stage of labor at 40 weeks' gestation shows successive episodes oflow (A-B) and high (B-C) FHR variation. The mean range oflong-term variation was 50 milliseconds, though episodes of low variation averaged 20.7 and high variation averaged 66 milliseconds. No analgesia was given; at delivery, umbilical arterial pH was 7,26 and base deficit was 7 mmol/L
In 73% of the records there were one or more episodes of both low and high FHR variation. In 23%, there was only one episode of high variation; this was not surprising because these episodes often exceed 1 hour in duration (see below). In 3%, there was only one episode of low variation. Amplitude of FHR variation in high or low episodes. First, we examined the amplitude of long-term FHR variation in early labor (~2 hours before the onset of the second stage), either spontaneous or induced. To obtain a fair sample, we selected patients whose fetuses were not acidemic at delivery, as judged by the umbilical arterial base deficit «8 mmoI/L), and for whom there were ~2 hours of data available. Table II shows no significant difference in the mean amplitude of FHR variation, in episodes of high or low variation, between 37 patients in whom labor was induced and 18 in whom it occurred spontaneously. For comparison, Table II shows that even in 15 fetuses who had severe metabolic acidemia at delivery (umbilical artery base deficit: range, 13 to 26 mmol/L; mean, 16.4 mmoIlL) ,
FHR variation was the same as in the other two groups at this early stage of labor. In 525 hours of data recorded for all 136 patients, there was no significant correlation between the FHR variation, recorded in episodes of high or low variation, and either gestational age (Fig. 2) or base deficit at delivery. The variation (mean ± SE) in low episodes was 20.6 ± 0.4 milliseconds (range 14.2 to 29.6 milliseconds); in high episodes it was 57.3 ± 1.1 milliseconds. Fig. 2 illustrates the large variation averaged per patient in early labor, from an extreme of 14.2 milliseconds (in low episodes) to 99.4 milliseconds (in high episodes). In the 91 patient records in which both were available, the ratio (mean ± SE) oflong-term FHR variation in episodes of low and high variation was calculated as 0, 39 ± 0 .008 with a range of 0.23 to 0.73. There was no correlation with base deficit at delivery, The ratio (low / high) was significantly (p < 0.02) related to the absolute value of FHR variation in episodes of low variation, but the correlation was low (r = 0.26). Hence, there was only a small tendency for those fetuses with
Episodic fetal heart rate variation
Volume 165 Number 3
Table II. Comparison of the amplitude of FHR variation in episodes of high or low variation
Induced labor «8 mmollL base deficit on delivery) Spontaneous labor «8 mmol/L base deficit on delivery) Labor ending in severe acidemia (> 12 mmoll L base deficit on delivery)
Amplitude of long-term FHR variation* (msec)
No. of patients
Mean hours per patient
56.0 ± 2.2
20.7 ± 0.6
58.4 ± 3.0
20.5 ± 0.6
56.7 ± 2.2
21.5 ± 1.0
Blood samples were taken from umbilical artery at delivery. *None of the mean differences are statistically significant.
lower FHR variation in low episodes to be lower in high episodes. Measurements of long-term FHR variation (in terms of pulse frequency, beats per minute) in episodes of low FHR variation are normally distributed (Fig. 3). The incidence of measurements of low amplitude «5 beats/min) was 11.1 % of the total (n = 287). Duration of episodes of high or low FHR variation. The duration (mean ± SE) of episodes of high FHR variation per hour was not significantly different in patients in spontaneous (19.9 ± 1.90 minutes) or induced (21.5 ± 1.14 minutes) labor. Nor was the duration (mean ± SE) of episodes of low FHR variation per hour different in spontaneous (14.5 ± l.10 minutes) or induced (12.9 ± 0.67 minutes) labor. Because the frequency distributions were not normal, these calculations were repeated with a log transformation, which normalized the distribution and gave the same results, that is, no significant differences. The durations per hour of high or low episodes of FHR variation were not significantly related to outcome as judged by the arterial base deficit at delivery. We concluded that it was safe to amalgamate the results from all fetuses. Episodes of low FHR variation were selected where bounded at either end by episodes of high variation and vice versa (see Fig. 1, A to B, B to C). Episodes of low variation may be interrupted transiently by a small brief deceleration or acceleration, and episodes of high variation by a brief deceleration; such transient interruptions were ignored for this purpose. There were 172 episodes of high variation in 85 patients, with a duration (mean ± SE) of 45.1 ± 2.7 minutes; there were 123 episodes of low variation in 62 patients, with a duration (mean ± SE) of 24.3 ± l.3 minutes. Fig. 4 shows the frequency distribution by duration. Episodes of high variation are usually longer, but even episodes of low variation sometimes exceed 60 minutes in early labor. There were four episodes of low variation (3.25%) lasting more than 60 minutes in four different patients (2.9%). Spencer andJohnson 8 reported the incidence of com-
plete cycles of low and high FHR variation (indicative of quiet or active fetal behavior) to be less common in spontaneous labor (38%) than after induction (68%) at term. Our figures showed an equal incidence in spontaneous (72.5%) and induced (72.7%) labors, judged by the presence of at least one complete cycle of both low and high FHR variation. Progress of labor and diurnal variation. To find whether the duration or amplitude of FHR variation (in high or low episodes) changed with time in early labor, we selected 51 patients for whom 5 to 12 hours of data were available. Thirty-three of the 51 patients had episodes of high variation, and 18 had episodes of low variation. The linear correlation of each variable (duration and amplitude of high and low variation) with time was calculated for each patient. In 7 of the 33 patients there was a significant (p < 0.05 to
HIGH EPISODES LOW EPISODES
Fig. 2. Distribution of the mean values of long-term FHR variation in episodes of high (e) or low (0) FHR variation with gestational age.
o a: o()
0 +-~-~---II, 0 1 2 3 4 5 6 7 8 9 10 11 12 13 FHR VARIATION (bpm) in LOW EPISODES
Fig. 3. Frequency distribution of long-term FHR variation (in terms of pulse frequ ency beats per minute) in episodes of low FHR variation. Eleven percent of these episodes had a mean value of