Fetal heart rate in relation to its variation in normal and growth retarded fetuses

European Journal of Obstetrics & Gynecology and Reproductive Biology 89 (2000) 27–33

www.elsevier.com / locate / ejogrb

Original Article

Fetal heart rate in relation to its variation in normal and growth retarded fetuses a, a a b c Ilse J.M. Nijhuis *, Judith ten Hof , Eduard J.H. Mulder , Jan G. Nijhuis , Harini Narayan , d a David J. Taylor , Gerard H.A. Visser a

b

Department of Obstetrics and Gynaecology, University Hospital, Utrecht, The Netherlands Department of Obstetrics and Gynaecology, University Hospital, Maastricht, The Netherlands c Department of Obstetrics and Gynaecology, Princess Margaret Hospital, Swindon, UK d Department of Obstetrics and Gynaecology, University of Leicester, UK Received 29 March 1999; received in revised form 2 June 1999; accepted 15 June 1999

Abstract Objectives: (1) to assess the relationship of basal fetal heart rate (FHR) with both long term (LTV) and short term (STV) FHR variation in low-risk pregnancies, longitudinally from 24 weeks gestation onwards and (2) to investigate the relationship of FHR with LTV and STV in intrauterine growth retarded (IUGR) fetuses. Study design: Computerised FHR recordings were made in twenty-nine uncomplicated pregnancies (n5224) and in twenty-seven IUGR fetuses who were selected retrospectively from three databases (n5135). Nomograms of FHR variation with FHR and GA were constructed using multilevel analysis. Results and conclusions: There was a strong negative relationship of FHR with both LTV and STV in the control group (R 2 553% and 52%, respectively). In the IUGR fetuses, FHR was generally higher than in normal fetuses whereas LTV and STV were lower. The relationship of FHR with LTV and STV in the IUGR group was less strong (for both: R 2 518%). Correction of FHR variation for basal FHR in the IUGR fetuses only resulted in a slight reduction in the number of recordings with a variation below the normal range. As it does not improve the recognition of fetuses being considered at the highest risk, such a correction of FHR variation for basal FHR is therefore not necessary. Intrafetal consistency, known to be present in healthy fetuses, was also present in the IUGR fetuses with a low FHR variation.  2000 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Human fetus; Fetal heart rate; Intrauterine growth-retardation; Intrafetal consistency; Longitudinal study design

1. Introduction Antenatal fetal heart rate (FHR) monitoring is widely used to assess the fetal condition. FHR variability is known to depend on several factors such as gestational age (GA), basal FHR, hypoxia, fetal behavioural states, cord compression, and the use of medication and social drugs (alcohol, smoking) [1]. Both long term (LTV) and short term (STV) FHR variation have a negative relationship *Corresponding author. Present address: Medisch Spectrum Twente, Department of Paediatrics, PO Box 50.000, 7500 KA Enschede, The Netherlands. Tel.: 131-53-487-2000 sign 235; fax: 131-53-487-2326. E-mail address: [email protected] (I.J.M. Nijhuis)

with FHR. FHR variation increases with decreasing FHR and vice versa. Moreover, FHR variation increases with advancing GA [1,2]. In intrauterine growth retarded (IUGR) fetuses, LTV is about 25% lower than in age-matched appropriately grown fetuses [3]. FHR variation gradually decreases with progressive compromise of the fetal condition [4]. This reduced FHR variation is not caused by a change in the rest-activity cycles [5]. However, it coincides with an increase in basal heart rate [4]. Usually around the same time that FHR variation falls below the norm, decelerations emerge [4,6]. Ribbert et al. [7] found LTV in IUGR fetuses to be reduced towards the lower level of the normal range about ten days before Caesarean Section (CS) was

0301-2115 / 00 / $ – see front matter  2000 Elsevier Science Ireland Ltd. All rights reserved. PII: S0301-2115( 99 )00162-1

28

I. J.M. Nijhuis et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 89 (2000) 27 – 33

undertaken for ‘fetal distress’. FHR variation remained fairly constant for some time before a further reduction occurred [7]. Decreased FHR variation and FHR decelerations have been found to be associated with hypoxaemia at CS and at cordocentesis [6,8–10]. Abnormal blood flow velocity waveforms from the umbilical artery are present in the majority of fetuses long before the occurrence of fetal distress [7,11]. The understanding of these temporal changes is important for the management of the IUGR fetus, especially with respect to the timing of delivery [12]. The negative relationship between basal heart rate and its variation in healthy fetuses, and the concomitant changes in rate (up) and variation (down) in IUGR fetuses, suggest that correction for rate may better identify those fetuses who have a reduced FHR variability irrespective of their basal heart rate. By studying FHR in relation to its variation, insight into changes in variation will improve. The aims of the present study were: (1) to assess the relationship between FHR and FHR variation (LTV and STV) in low-risk pregnancies longitudinally from 24 weeks gestation onwards; and (2) to investigate the same relationship in the IUGR fetus.

2. Subjects and methods FHR was monitored longitudinally in 29 singleton fetuses of healthy pregnant women, of which eleven were nulliparous. From 24 weeks GA onwards one hour recordings were made at two-weekly intervals. From 36 weeks till delivery, two hour recordings were made on a weekly basis, but only the first hour was used for the present analysis. The recordings took place at the Leicester General Hospital, Leicester, UK. All women gave their written informed consent. All pregnancies were uncomplicated, and no medication was given. The pregnancies were reliably dated using the last menstrual period or by ultrasound scan in cases of uncertain dates. An anomaly scan performed in all cases at 18–20 weeks detected no structural abnormalities. All children were born after 36 weeks GA (median 39 2 / 7 ; range 36 1 / 7 –41 6 / 7 ) with a birthweight .10th centile for GA according to customised growth charts [13] (median 3175 g; range 2665–4500 g). All infants were healthy at birth except one girl who had an opacity of the eye lenses (Peters’ anomaly [14]). She will be visually handicapped. She was not excluded from the study on the assumption that this isolated abnormality would not affect fetal heart rate. Separate analysis of her data showed that all FHR parameters were well within the normal range. Neurological examinations at three months of age revealed no abnormalities, except for the eye anomaly. Nomograms of FHR and FHR variation have already been developed out of this control group [15]. The IUGR fetuses were selected retrospectively from three Dutch databases, using the following criteria: (I)

clinical diagnosis of IUGR with ultrasound measurements ,5th centile; (II) birthweight ,10th centile for GA corrected for sex and parity [16]; (III) $5 computerised FHR recordings made on different days; (IV) CS performed for suspected ‘fetal distress’ (repetitive decelerative or terminal FHR pattern [17] and / or abnormal pulsatility index (PI) of the umbilical artery with decreasing FHR variation); and (V) no congenital abnormalities. In total 27 fetuses fulfilled all five criteria and were thus selected. Of these, 12 have been described previously [7]. In these 27 IUGR fetuses, FHR variation had been investigated in 14 fetuses by measuring LTV, and in the other 13 by measuring STV. The median GA at birth was 31 5 / 7 (range 29 0 / 7 –36 0 / 7 ) and median birthweight was 1010 g (range 640–1580 g). The pH of the umbilical artery was measured in 24 fetuses (the median pH was 7.21; range 6.88–7.30). The fetuses were monitored over a median period of 17 days (range 5–44 days). The PI of the umbilical artery was .2SD in 25 fetuses; oligohydramnios was present in 11 fetuses. Fifteen mothers received antihypertensive drugs (amongst others methyldopa, labetalol, hydralazine and magnesium sulphate) and / or low dose aspirin. Seven women received betamethasone to enhance fetal lung maturation. FHR was monitored with the women lying in a semirecumbent position using the Sonicaid System 8000  or 8002  (Oxford, Sonicaid Ltd. Chichester, UK). This system reduces the FHR data over 3.75 s (1 / 16 min epochs). Basal FHR (in beats per minute (bpm)) is measured as the mean rate averaged over all periods of low variation or, if no low variation is present, it is derived from a statistical analysis. LTV is calculated as the average of 1 min ranges of pulse intervals about a computed baseline, excluding data from decelerations and errors (‘mean minute range’, in ms). STV, not measured by the System 8000  we used, is the average of sequential 1 / 16th min pulse interval differences, after exclusion of FHR decelerations. Signal loss is calculated as a percentage of the whole recording. Signal loss of 30% or more is noted by the System as a high loss and was also used as an exclusion criterion in our study [18]. FHR variation was preferably assessed by STV, but LTV was used when STV was not available. The correlation between FHR and FHR variation was calculated in both the control and case groups using Spearman correlation coefficient. Multilevel analysis [19], a recently developed statistical technique, was carried out with the software program Mln (Multilevel Model Project, London, UK) to calculate, from 24 weeks GA onwards, the reference ranges of LTV and STV corrected for basal FHR. This method has been recently used by us to develop nomograms for basal FHR, LTV and STV out of 60 min recordings [15]. For the present analysis, we tested the influence of fetal sex on the intercept and regression coefficient. The effect of time of the day at which the recording was made and the effect of

I. J.M. Nijhuis et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 89 (2000) 27 – 33

FHR on FHR variation were tested on the intercept. Both linear and quadratic functions were tested. Five recordings were randomly selected for analysis from each IUGR fetus. These values were plotted in two different reference curves of the control group: (1) the three-dimensional graphs corrected for basal FHR as developed in this article and (2) the two-dimensional nomograms of FHR variation with GA, not corrected for FHR [15]. The number of values falling below the 2.5th centile in the two different graphs was counted for each fetus separately. To test whether the use of the two- or three-dimensional graph lead to a different result, i.e. more or less values below the 2.5th centile, the Wilcoxon

29

matched-pairs signed-rank test was used to compare the numbers of abnormal values of FHR variation in the graphs corrected and not corrected for basal FHR, respectively.

3. Results We analysed a total of 224 recordings from the 29 fetuses of the control group, after exclusion of 14 recordings with a signal loss .30%. FHR was significantly correlated with both LTV and STV. The following models were derived; for LTV: y 5 2 1.15x 1 208.7 (R50.73;

Fig. 1. Three-dimensional nomogram of long term FHR variation (LTV) with basal FHR and gestational age. LTV increases during gestation and decreases with increasing FHR.

30

I. J.M. Nijhuis et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 89 (2000) 27 – 33

p,0.0001, n5224) and for STV: y 5 2 0.21x 1 37.7 (R50.72; p ,0.0001, n5224), where y is LTV (ms) or STV (ms) and x is basal FHR (bpm)). Multilevel analysis confirmed this strong relationship between FHR and both LTV and STV ( p,0.0001) during gestation. For both LTV and STV, the relative goodness of fit significantly improved by adding a quadratic component of GA to the model. After correction for basal FHR, no significant effect of time of day or fetal sex on LTV or STV was found. Three dimensional charts were made to show the relation of LTV and STV with GA and basal FHR (Figs. 1 and 2, respectively). The coefficients used to construct these figures are given in Table 1.

In the 27 IUGR fetuses, a total of 340 recordings was made, with a median of 12 recordings per fetus (range 5–30). To compare values plotted in the graphs corrected and uncorrected for FHR, we analysed 5 recordings per fetus (chosen at random), most of which had a duration of 60 min. Eighteen of these 135 recordings, all made before 31 weeks GA, were of 30–40 min duration. This is long enough to reliably assess FHR variation at this age [15]. Of the 135 recordings, 70 recordings (of 14 fetuses) were analysed using LTV and 65 (of 13 fetuses) using STV. FHR (bpm) was significantly correlated with both LTV (ms) and STV (ms) ( y 5 2 0.37x 1 80.3; R50.42; p,0.0001, n5 70 and y 5 2 0.13x 1 24.5; R50.43; p,0.0001, n565,

Fig. 2. Three-dimensional nomogram of short term FHR variation (STV) with basal FHR and gestational age. STV increases during gestation and decreases with increasing FHR.

I. J.M. Nijhuis et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 89 (2000) 27 – 33

31

Table 1 Coefficients used for the graphs of LTV and STV with basal FHR and gestational age in weeks (WGA)a y

b

a1

a2

a3

LTV (ms) STV (ms) Variance LTV Variance STV

151.8 28.07 21.32 1.11

20.80 20.15 – –

1.73 0.45 4.76 0.13

20.066 20.022 0.185 0.009

a The equations must be read as y 5 b 1 a 1 (FHR) 1 a 2 (WGA 2 24) 1 a 3 (WGA 2 24)2 .

respectively), but the correlation coefficients were lower than in the low-risk pregnancies. In the IUGR fetuses, FHR was generally higher than in normal fetuses (in 99 of 135 records (73%) FHR was higher than the 50th centile), whereas FHR variation was lower, with 131 records (97%) below the 50th centile (Figs. 3 (LTV) and 4 (STV)). Deviation from the median and the number of records with values exceeding the 97.5th centile (FHR) or below the 2.5th centile (FHR variation) were much higher for variation than for rate (n572 and n55, respectively). Correction of LTV and STV for FHR, by plotting the values in Figs. 1 or 2, respectively, resulted only in a slight reduction in the Fig. 4. Five individual data points of 13 intrauterine growth retarded fetuses plotted in nomograms of baseline fetal heart rate (bpm) and short term FHR variation (ms) with gestational age. Given are the P50 (solid line), P2.5 and P97.5 (dashed lines) of the control group.

number of recordings with a variation below the 2.5th centile (n557 as compared to n572; p50.12). Intrafetal consistency, known to be present in healthy fetuses, was also present in the IUGR fetuses. Although this can be seen in Figs. 3 and 4, it is better illustrated by five subjects of whom all available data points were plotted (Fig. 5).

4. Comment

Fig. 3. Five individual data points of 14 intrauterine growth retarded fetuses plotted in nomograms of baseline fetal heart rate (bpm) and long term FHR variation (ms) with gestational age. Given are the P50 (solid line), P2.5 and P97.5 (dashed lines) of the control group.

We have previously shown that FHR variation is negatively correlated with basal heart rate in healthy fetuses, both during ‘non-reactive’ FHR patterns (pattern A) and ‘reactive’ FHR patterns (pattern B) [20]. In the present study this was confirmed when the whole one hour recordings were considered. The relationship was strong and about 50% of the differences in FHR variation could be explained by differences in rate (R 2 553% for LTV and R 2 552% for STV, respectively). This implies that basal heart rate is a strong determinant of heart rate variation, the more so since there are large interfetal differences in rate. FHR variation increases in the course of the day [15,21]. This increase can be explained by a decrease in rate, since correction for rate resulted in absence of an effect of time

32

I. J.M. Nijhuis et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 89 (2000) 27 – 33

Fig. 5. The existence of intrafetal consistency in intrauterine growth retarded fetuses is illustrated in five fetuses. All available individual data points are plotted in nomograms of baseline fetal heart rate (bpm) and short term FHR variation (ms) with gestational age. Given are the P50 (solid line), P2.5 and P97.5 (dashed lines) of the control group.

of the day on LTV and STV. Our study was restricted to the day-time. With respect to diurnal rhythms it is unlikely that changes in rate can explain all 24 h changes in FHR variation [21]. In the IUGR fetuses, FHR was generally higher and FHR variation was lower than in the control group. The increase in heart rate has been described before [4,6,8] and is thought to occur at about the same time as FHR decelerations occur [4]; in other words it may be an adaptation to a hypoxaemic condition. Part of the reduction in FHR variation may be explained by the increase in basal heart rate. However, correction for rate did not result in large changes in the proportion of records with a FHR variation below the normal range. This may be explained by the fact that changes in FHR variation exceeded those in basal heart rate (Figs. 3 and 4). Correction for basal heart rate will therefore not result in significant changes in the population identified as being at the highest risk, and seems therefore not necessary. The correlation between FHR and its variation was lower in the IUGR fetuses than in the healthy fetuses. This

may be due to the fact that the relationship between these two variables depends on the actual fetal condition. Initially, FHR will increase whereas FHR variation falls (negative correlation). However, with further deterioration of the fetal condition, FHR variation will decrease progressively whereas basal heart rate does not increase further and may even return to normal values (‘terminal FHR pattern’ [17,22]). Under such circumstances a positive correlation between heart rate and its variation will be present. In this study we could not investigate this in detail, given the retrospective study design and the use of various drugs which may also affect FHR and its variation. For instance, hydralazine and magnesium sulphate may increase basal heart rate, although such an increase was also found with increasing severity of maternal hypertension when these drugs were not given [23]. Methyldopa was found to have no effect on FHR [24], while FHR decreased after labetalol was given to the mother [25]. Betamethasone temporarily decreases FHR variation and this decrease cannot be explained by changes in basal heart rate [26]. However, when the data were analysed using only the fetuses not under influence of medication similar results were obtained (data not shown). So, changes in FHR variability can, although for a large part, being explained by changes in fetal heart rate. The clinical implications of (changes in) FHR variability, as also described in the introduction, are not yet fully clear, but a decrease in FHR variability without an increase in basal FHR might be a sign for extreme deterioration of the fetus which needs immediate intervention. To investigate this in more detail, future prospective studies are needed. In low-risk fetuses, the data of individual fetuses show a homogeneous trend and level with large interfetal differences with respect to the individually measured values. Individual low-risk fetuses ‘use’ only about 40% of the total bandwidth of the normal range [15]. This intrafetal consistency seems also to be present in IUGR fetuses even when FHR variation falls below the norm (Figs. 3–5). As stressed by others, it is therefore important to use each fetus as its own control, especially for monitoring of trends and detection of minor changes [4,15,21,27,28]. The background of the large interfetal differences in heart rate (in healthy fetuses) has been further examined in another study (ten Hof J, Nijhuis IJM, Mulder EJH et al. Fetal and maternal heart rate parameters: their relationship during the second half of gestation and effects of maternal stress. Submitted). In conclusion, there is a strong relationship between FHR and both LTV and STV. This correlation is greater for healthy fetuses than for IUGR fetuses. Correction for basal heart rate does not result in a significant change in the number of recordings which are identified as having a FHR variation below the normal range. Correction of FHR variation for basal FHR is therefore not necessary in IUGR fetuses.

I. J.M. Nijhuis et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 89 (2000) 27 – 33

Acknowledgements The authors like to thank L.S.M. Ribbert, MD PhD (Department of Obstetrics and Gynaecology, St. Antonius Hospital Nieuwegein, The Netherlands), S.V. Koenen, MD and R.H. Stigter, MD (Department of Obstetrics and Gynaecology, University Hospital Utrecht, The Netherlands) for the use of their data, and P. Westers, PhD (Centre for Biostatistics, Utrecht University, The Netherlands) for his statistical advice. This study was supported by a grant from the Commission of the European Communities (Contract No. CHRXCT 94-0613).

References [1] van Geijn HP. Developments in CTG analysis. Bailliere’s Clin Obstet Gynaecol 1996;10(2):185–209. [2] Martin Jr. CB. Regulation of the fetal heart rate and genesis of FHR patterns. Semin Perinatol 1978;2:131–46. [3] Gagnon R, Hunse C, Fellows F, Carmichael L, Patrick J. Fetal heart rate and activity patterns in growth-retarded fetuses: changes after vibratory acoustic stimulation. Am J Obstet Gynecol 1988;158:265– 71. [4] Snijders RJM, Ribbert LSM, Visser GHA, Mulder EJH. Numeric analysis of heart rate variation in intrauterine growth-retarded fetuses: a longitudinal study. Am J Obstet Gynecol 1992;166:22–7. [5] Henson G, Dawes GS, Redman CWG. Characterization of the reduced heart rate variation in growth-retarded fetuses. Br J Obstet Gynaecol 1984;91:751–5. [6] Bekedam DJ, Visser GHA, Mulder EJH, Poelmann-Weesjes G. Heart rate variation and movement incidence in growth-retarded fetuses: the significance of antenatal late heart rate decelerations. Am J Obstet Gynecol 1987;157:126–33. [7] Ribbert LSM, Visser GHA, Mulder EJH, Zonneveld MF, Morssink LP. Changes with time in fetal heart rate variation, movement incidences and haemodynamics in intrauterine growth retarded fetuses: a longitudinal approach to the assessment of fetal wellbeing. Early Hum Dev 1993;31:195–208. [8] Visser GHA, Sadovsky G, Nicolaides KH. Antepartum heart rate patterns in small-for-gestational-age third-trimester fetuses: correlations with blood gas values obtained at cordocentesis. Am J Obstet Gynecol 1990;162:698–703. [9] Smith JH, Anand KJS, Cotes PM, Dawes GS, Harkness RA, Howlett TA et al. Antenatal fetal heart rate variation in relation to the respiratory and metabolic status of the compromised human fetus. Br J Obstet Gynaecol 1988;95:980–9. [10] Ribbert LSM, Snijders RJM, Nicolaides KH, Visser GHA. Relation of fetal blood gases and data from computer-assisted analysis of fetal heart rate patterns in small for gestation fetuses. Br J Obstet Gynaecol 1991;98:820–3. [11] Bekedam DJ, Visser GHA, van der Zee AGJ, Snijders RJM,

[12]

[13]

[14] [15]

[16] [17] [18] [19] [20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

33

Poelmann Weesjes G. Abnormal velocity waveforms of the umbilical artery in growth retarded fetuses: relationship to antepartum late heart rate decelerations and outcome. Early Hum Dev 1990;24:79– 89. Visser GHA. Assessment of fetal well-being in growth-retarded fetuses. In: Hanson MA, Spencer JAD, Rodeck CH, editors, Fetus and Neonate. Physiological and clinical applications, (vol. 3 Growth), Cambridge: Cambridge University Press, 1995, pp. 327– 45. Gardosi J, Mongelli M, Wilcox M, Chang A, Sahota D, Francis A. Gestation related optimal weight (GROW) program. Software version 2, PRAM, Nottingham University, 1997. Wright KW. Pediatric ophthalmology and strabismus, St. Louis: Mosby, 1995. Nijhuis IJM, ten Hof J, Mulder EJH, Nijhuis JG, Narayan H, Taylor DJ et al. Numerical fetal heart rate analysis: nomograms, minimal duration of recording and intrafetal consistency. Prenat Neonat Med 1998;3(3):314–22. Kloosterman GJ. On intrauterine growth, Int. J Gynaecol Obst 1970;8:895–912. Visser GHA, Huisjes HJ. Diagnostic value of the unstressed antepartum cardiotocogram. Br J Obstet Gynaecol 1977;84:321–6. Dawes GS, Moulden M, Redman CWG. System 8000: computerized antenatal FHR analysis. J Perinat Med 1991;19:47–51. Goldstein H. Multilevel statistical models, 2nd ed, London: University of London, 1995. Nijhuis IJM, ten Hof J, Mulder EJH, Nijhuis JG, Narayan H, Taylor DJ. et al. Fetal Heart Rate (FHR) parameters during FHR patterns A and B: a longitudinal study from 24 weeks’ gestation. Prenat Neonat Med 1998;3(4):383–93. Visser GHA, Dawes GS, Redman CWG. Numerical analysis of the normal human antenatal fetal heart rate. Br J Obstet Gynaecol 1981;88:792–802. Visser GHA, Redman CWG, Huisjes HJ, Turnbull AC. Nonstressed antepartum heart rate monitoring: implications of decelerations after spontaneous contractions. Am J Obstet Gynecol 1980;138:429–35. Lenox JW, Uguru V, Cibils LA. Effects of hypertension on pregnancy monitoring and results. Am J Obstet Gynecol 1990;163:1173–9. Wide-Swensson D, Montan S, Arulkumaran S, Ingemarsson I, Ratnam SS. Effect of methyldopa and isradipine on fetal heart rate pattern assessed by computerized cardiotocography in human pregnancy. Am J Obstet Gynecol 1993;169:1581–5. Harper A, Murnaghan GA. Maternal and fetal haemodynamics in hypertensive pregnancies during maternal treatment with intravenous hydralazine or labetalol. Br J Obstet Gynaecol 1991;98:453–9. Mulder EJH, Derks JB, Visser GHA. Antenatal corticosteroid therapy and fetal behaviour: a randomised study of the effects of betamethasone and dexamethasone. Br J Obstet Gynecol 1997;104:1239–47. Smith JH, Dawes GS, Redman CWG. Low human fetal heart rate variation in normal pregnancy. Br J Obstet Gynaecol 1987;94:656– 64. Visser GHA, Mulder EJH, Stevens H, Verweij R. Heart rate variation during fetal behavioural states 1 and 2. Early Hum Dev 1993;34:21–8.