Maternal and neonatal effects of bolus administration of ephedrine and phenylephrine during spinal anaesthesia for caesarean delivery: a randomised study

Introduction 

Hypotension, a frequent complication of spinal anaesthesia for caesarean delivery, can have adverse maternal (nausea, vomiting, dizziness, and decreased consciousness) and fetal effects (decreased uteroplacental blood flow, impaired fetal oxygenation, and fetal acidosis).1 Vasopressors are often required and current evidence suggests that phenylephrine is more suitable than ephedrine in obstetrics.2 The propensity of ephedrine to cause maternal tachycardia and depress fetal pH and base excess has resulted in suggestions to eliminate or drastically limit its use.3 Most clinical trials have focused on relatively high-dose prophylactic infusions of vasopressor.[4], [5], [6], [7], [8] Outcome may possibly be different if low-dose intermittent boluses of ephedrine are used to treat hypotension. Indeed three previous studies comparing boluses of ephedrine and phenylephrine for post spinal hypotension found no significant differences in efficacy or in the incidence of fetal acidosis.[9], [10], [11]

In studies comparing ephedrine and phenylephrine, assessment of the newborn has been limited to Apgar scores and umbilical acid-base status. Neurobehavioural testing, used elsewhere in assessment of the baby,12 has not been included when comparing vasopressors. We therefore included neurobehavioural assessment in a comparison of ephedrine and phenylephrine administered as intermittent boluses to treat post spinal hypotension. We postulated that any advantage of phenylephrine over ephedrine in terms of improved neonatal acid-base status would probably be reflected in the neurobehavioural response scores of the newborn.

Methods 

After obtaining approval from the hospital ethics committee and informed written consent, 60 ASA I women with singleton pregnancies scheduled for elective caesarean delivery under spinal anaesthesia were recruited in this randomised and double-blind study. Women with pre-existing or pregnancy-induced hypertension, diabetes mellitus, known cardiovascular or cerebrovascular disease, fetal abnormality, or contraindication to spinal anaesthesia were excluded.

All women received ranitidine and metoclopramide antacid prophylaxis. In the operating theatre, standard monitoring with non-invasive arterial pressure, electrocardiography and pulse oximetry was established. Women rested undisturbed in the supine position with left uterine displacement for 5min, following which baseline blood pressure and heart rate were calculated as the mean of three successive readings measured 1min apart. An 18-gauge intravenous cannula was sited. Each patient received a 10-mL/kg i.v. infusion of Ringer’s lactate solution over 15-20min before spinal anaesthesia. After prehydration, the fluid infusion was continued at minimal rate to maintain vein patency, regardless of any maternal haemodynamic changes. With the patient in the left lateral position, 2mL of 0.5% hyperbaric bupivacaine was injected intrathecally at L3-4 via a 25-gauge Quincke spinal needle. Patients were then immediately turned supine and positioned with left uterine displacement. Heart rate and blood pressure were recorded at 1-min intervals from the time of induction of spinal anaesthesia until delivery. Oxygen, 6 L/min was delivered via a face mask until delivery. The dermatomal level of anaesthesia, assessed by loss of pin prick discrimination, was recorded 5 and 15min after induction of spinal anaesthesia. Sensory block to T5 dermatome was considered adequate for surgery.

Women were randomly assigned to receive one of two vasopressor solutions whenever maternal systolic pressure decreased to 80% of baseline or less. Group E received a 1-mL bolus of ephedrine 6mg/mL; group P a 1-mL bolus of phenylephrine 100μg/mL. Additional boluses were administered if the systolic pressure remained at or below 80% of baseline. Women were randomised by computer-generated number allocation. The study drugs were prepared in identical 10-mL syringes by an anaesthesiologist not involved with data collection.

Atropine was administered in 0.3-mg increments whenever bradycardia (heart rate <60 beats/min) was associated with a systolic pressure less than baseline or if the heart rate was <45 beats/min irrespective of arterial pressure. The incidence of maternal tachycardia (heart rate >100 beats/min) and reactive hypertension (increase in systolic pressure above baseline by 20%) were recorded. The number of vasopressor doses required, total dose of vasopressor administered, time of first administration of vasopressor, requirement for atropine and its relation to vasopressor administration were noted.

The time of induction of spinal anaesthesia, uterine incision and delivery were recorded. After delivery oxytocin 5 units was given by slow i.v. injection followed by a 10-unit infusion. The incidence of nausea and vomiting was recorded. Arterial and venous blood samples were obtained from a double-clamped segment of umbilical cord and analysed within 10min. Apgar scores at 1, 5 and 10min were determined by the attending paediatrician who was unaware of group assignment. Time of onset of sustained rhythmic respiration was noted. Neurobehavioural response of the neonate was assessed at 2-4h, 24±2h, and at 48±2h of age by a modified Early Neonatal Neurobehavioral Scale (ENNS).12 This examination involved an assessment of neonatal reflexes, response to stimuli such as sound and pinprick, evaluation of general body and truncal tone, and general alertness. Absent and weak responses were grouped as low scores, and moderate and brisk responses as high scores. The anaesthesiologist assessing the neurobehavioural response was blind to group allocation.

Results 

Ninety-eight women were enrolled in the study (Fig. 1); 38 did not develop hypotension (systolic pressure ⩽80% of baseline) and did not require vasopressors. These subjects were excluded from the study. The randomisation code was not broken and further subjects were recruited. A total of 60 women who developed hypotension participated in the study. Due to technical difficulties umbilical artery blood samples could not be obtained from one subject in each group.

View Large Image Download to PowerPoint

Fig. 1. Flow of participants in the randomised trial. ENNS: early neonatal neurobehavioural scale.

Physical characteristics, baseline haemodynamic data and dermatomal sensory levels are presented in Table 1. The two groups were comparable in age, weight, height, baseline haemodynamic data and dermatomal sensory levels. There were no significant differences in the mean induction-to-delivery or uterine-to-delivery intervals between the two groups.

Table 1. Patient characteristics, baseline haemodynamic data, dermatomal anaesthetic levels and surgical times

		Ephedrine (n=30)	Phenylephrine (n=30)
	Age (years)	24.5±4.4	24.1±4.4
	Weight (kg)	56.6±6.7	58.7±5.6
	Height (cm)	154.6±5.9	155.2±5.7
	Baseline systolic pressure (mmHg)	128±8	126±9
	Baseline heart rate (beats/min)	86±7	86±7
	Block height at 5min	T5 (T4-T8)	T5 (T4-T7)
	Block height at 15min	T4 (T3-T5)	T4 (T3-T5)
	Induction-delivery time (min)	13.4±2.8	15.4±3.9
	Uterine incision-delivery time (s)	46±19	53±18

Values are mean±SD or median (range). No significant differences between groups.

Maternal haemodynamic data are summarized in Table 2. Changes in systolic pressure and heart rate are shown in Fig. 2, Fig. 3. As the time from induction to delivery varied among patients, haemodynamic changes were compared up to 20min after induction of spinal anaesthesia, by which time 59 of 60 women had delivered. The mean change in systolic pressure was comparable in the two groups, except at time-point T8 (P=0.004). The fall in heart rate below mean baseline in group P was significantly greater than that in group E (20±10 vs. 6±0.6, respectively; P<0.001). In all cases, bradycardia developed following phenylephrine administration. Four women in group E (13%) compared with none in group P reported nausea (P>0.05). One woman in group E vomited (P=0.313). No woman in either group complained of dizziness.

Table 2. Maternal haemodynamic data

		Ephedrine (n=30)	Phenylephrine (n=30)	P
	Heart rate <60 beats/min	0 (0)	5 (16.7)	0.052
	Bradycardia+hypotension	0 (0)	0 (0)
	Heart rate >100 beats/min	13 (43)	6 (20)	0.095
	Number of vasopressor doses	2 (1-4)	2 (1-3)	0.152
	Total vasopressor dose (mg)	12.5±5.1	0.16±0.06
	Time to first vasopressor (min)	5.5±1.9	6.0±1.9	0.314
	Minimum systolic pressure (mmHg)	100 (80-112)	93 (70-110)	0.114
	Maximum systolic pressure (mmHg)	122 (110-140)	127 (100-150)	0.059
	Reactive hypertension	0 (0)	0 (0)
	Atropine requirement	0 (0)	3 (10)	0.206

Values are number (%), mean±SD, or median (range).

View Large Image Download to PowerPoint

Fig. 2. Serial changes in mean systolic pressure. Changes over time were not significantly different between ephedrine (white circles) and phenylephrine (black circles) groups, except at time-point T8 (P=0.004). Vertical bars represent standard deviations.

View Large Image Download to PowerPoint

Fig. 3. Serial changes in mean heart rate. Changes over time were significantly different between ephedrine (white circles) and phenylephrine (black circles) groups between time-points T6 to T19 (P<0.05). Vertical bars represent standard deviations.

Neonatal data are presented in Table 3. Birth weight and Apgar scores at 1, 5 and 10min were comparable in the two groups. No neonate had an Apgar score <7 at any time point. Time to onset of sustained rhythmic respiration was <90 s in all cases. No neonate required tracheal intubation or admission to the neonatal intensive care unit. No umbilical artery pH values were less than 7.20. Umbilical artery and venous pH were significantly lower in group E than in group P (P=0.01 and P=0.002, respectively). Umbilical artery and venous base deficit were significantly greater in group E than in group P (P<0.001 and P=0.001, respectively). Neurobehavioral scores at 2-4h, 24h and 48h were comparable in the two groups (Table 4).

Table 3. Neonatal data

		Ephedrine (n=30)	Phenylephrine (n=30)	P
	Birth weight (kg)	2.84±0.40	2.88±0.32	0.648
	Apgar scores at 1min	8 (7-9)	9 (8-9)	0.739
	5min	9 (7-10)	10 (8-10)	0.128
	10min	9 (8-10)	10 (8-10)	0.611
	Umbilical arterial acid-base status
	pH	7.29±0.04	7.32±0.04	0.010
	PO2 (mmHg)	17.5±2.83	18.13±3.21	0.459
	PCO2 (mmHg)	44.0±6.13	43.50±5.29	0.751
	Base excess (mEq/L)	-2.83±0.94	-1.61±1.04	<0.001
	Umbilical venous acid-base status
	pH	7.34±0.04	7.38±0.05	0.002
	PO2 (mmHg)	26.5±3.24	28.1±3.62	0.079
	PCO2 (mmHg)	35.9±5.45	36.0±4.72	0.939
	Base excess (mEq/L)	-1.9±0.76	-1.1±1.12	0.001

Values are mean±SD or median (range).

Table 4. Numbers of infants with low neurobehavioral scores at 2-4h, 24h, and 48h

		2-4h		24h		48h
		Ephedrine (n=30)	Phenylephrine (n=30)	Ephedrine (n=30)	Phenylephrine (n=30)	Ephedrine (n=30)	Phenylephrine (n=30)
	Response to pinprick	5 (17)	4 (13)	2 (7)	2 (7)	0 (0)	0 (0)
	Arm recoil	4 (13)	4 (13)	0 (0)	1 (3)	0 (0)	0 (0)
	General body tone	8 (27)	7 (23)	3 (10)	3 (10)	2 (7)	1 (3)
	Truncal tone	7 (23)	7 (23)	3 (10)	3 (10)	2 (7)	1 (3)
	Rooting	6 (20)	5 (17)	0 (0)	0 (0)	0 (0)	0 (0)
	Sucking	5 (17)	4 (13)	0 (0)	0 (0)	0 (0)	0 (0)
	Moro’s reflex	6 (20)	7 (23)	0 (0)	0 (0)	0 (0)	0 (0)
	Response to sound	6 (20)	8 (27)	0 (0)	0 (0)	0 (0)	0 (0)
	Placing	4 (13)	4 (13)	0 (0)	0 (0)	0 (0)	0 (0)
	Alertness	2 (7)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)

Values are number (percent). No significant differences between groups.

Discussion 

Our results confirm that 100-μg bolus doses of phenylephrine are as effective as 6-mg bolus doses of ephedrine in the treatment of hypotension following spinal anaesthesia in term parturients undergoing caesarean delivery. The neonates of women treated with phenylephrine had higher umbilical cord pH and base deficit values though fetal acidosis (umbilical artery pH <7.20) was not seen in any neonate. Apgar scores at 1, 5, and 10min, and neurobehavioural scores at 2-4h, 24h and 48h after birth were similar between the two groups.

In a systematic review of randomised controlled trials, Lee et al.13 showed that ephedrine and phenylephrine have similar efficacy for preventing or treating hypotension, and although ephedrine use was associated with lower umbilical cord blood pH values than phenylephrine, there was no difference in Apgar scores. Women receiving phenylephrine are more likely to develop bradycardia than those treated with ephedrine.[11], [13] In our study 17% of women receiving phenylephrine developed a bradycardia of less than 60 beats/min compared with none in the ephedrine group. This difference was not statistically significant probably because the sample size was insufficient. It is noteworthy that no woman developed bradycardia with hypotension. In all cases bradycardia developed following phenylephrine administration, suggesting a baroreceptor reflex mechanism.

Comparative studies suggest that the use of phenylephrine is associated with better fetal acid-base status.[9], [10], [13] Although we found a statistically significant difference between groups in umbilical artery and vein pH, the differences were small and all values were within normal range for neonates. The cause for decreased umbilical cord pH with ephedrine is unclear. Blood pressure control was similar with ephedrine and phenylephrine in our study, suggesting that factors other than maternal hypotension are involved. Cooper et al.8 suggest increased fetal metabolic rate secondary to ephedrine-induced β-adrenergic stimulation, as the probable mechanism for increased incidence of fetal acidosis with ephedrine. With ephedrine, but not phenylephrine, decreasing umbilical artery pH was associated with an increasing umbilical arterial-venous PCO2 difference, suggesting an increased fetal metabolic rate.14

An interesting finding of our study is the lack of fetal acidosis (UA pH <7.20) in the ephedrine group. Protocols aimed at preventing hypotension have been considered to result in better neonatal outcome than those in which hypotension is treated after it has occurred.15 However, prophylactic administration of vasopressors is associated with the use of relatively large doses. Furthermore the use of prophylactic ephedrine has been associated with a relatively high incidence of fetal acidosis.[5], [7], [8] A meta-analysis focusing on the use of prophylactic ephedrine during spinal anaesthesia for elective caesarean delivery found a significant dose response relationship for hypotension, hypertension, and UA pH but not for nausea or vomiting, fetal acidosis or Apgar scores.16 Recently, Cooper et al.17 reported a 50% incidence of fetal acidosis with prophylactic ephedrine (mean total dose 43mg) in elective caesarean delivery. Contrary to this, the incidence of fetal acidosis was only 17% in a higher risk population having urgent caesarean delivery under spinal anaesthesia where ephedrine was given as required, rather than prophylactically.18 Likewise, neonatal outcome was similar between groups when relatively small doses of phenylephrine or ephedrine were used to treat hypotension in patients having non-elective caesarean delivery under spinal anaesthesia.19 In our study, the as-required use of ephedrine and the short spinal-to-delivery interval may have accounted for the lack of fetal acidosis in the ephedrine group by decreasing both the total dose and duration of fetal exposure to ephedrine.

Neurobehavioural assessment, though not universally accepted, has been used to examine the effects of intrapartum medication on the newborn. The anaesthesiologist assessing the neurobehavioural scores was trained by a paediatrician before the start of the study. So that inter-observer variation could be eliminated, only one evaluator, blind to group allocation, assessed neurobehavioural responses. Despite widespread use of the Neurologic and Adaptive Capacity Score (NACS),20 the reliability and validity of this score has not been established.21 The Brazelton neonatal neurobehavioral assessment scale,22 a more thorough test, has the disadvantage of requiring 45min to complete even when performed by an experienced person. Scanlon et al.12 devised a simple, less time consuming test, the ENNS, by adapting different elements from the Prechtal-Beintema neurological examination as well as the Brazelton scale.

In women undergoing caesarean delivery under spinal anaesthesia, neurobehavioural scores could be affected by placental transfer of vasopressor or by maternal hypotension resulting in fetal acidosis. Hughes et al.23 showed that ephedrine readily crosses the placenta, with an umbilical vein: maternal artery ratio of 0.71. No difference has been demonstrated in neurobehavioural scores if the duration of maternal hypotension is less than 2min.24 In our study, the percentage of neonates with low neurobehavioural scores were similar in the two groups at 2-4h, 24h and 48h after delivery. In both groups neurobehavioural scores improved with time, with all neonates demonstrating high scores for most of the variables at 48h. Early detection and prompt treatment of maternal hypotension with ephedrine or phenylephrine and the use of relatively low doses of vasopressor in the present study could explain the absence of significant differences in neurobehavioural scores between groups.

Our study has a number of limitations. First, the sample size was based on UA pH and was not powered to detect a significant difference in neurobehavioural scores. The similarity in neurobehavioural scores may therefore reflect a type-II error. Our results were obtained from healthy women undergoing elective caesarean delivery, so extrapolation to situations where fetal compromise is present or to emergency caesarean delivery may not be valid. The small decrease in pH that occurs with ephedrine may not be tolerated by such at-risk babies. In a chronic sheep model of increased placental vascular resistance, phenylephrine administration was associated with impaired uterine and placental haemodynamics and increased fetal lactate concentrations compared with ephedrine administration.25 Whether phenylephrine is superior to ephedrine in a compromised human maternal-fetal unit needs to be addressed. Ngan Kee et al.19 have recently provided evidence suggesting that phenylephrine is an appropriate vasopressor in non-elective cases.

The baseline blood pressure measurements were obtained in the operating theatre. This relatively stressful environment may explain the high baseline pressures seen in our population. All our patients received a fluid preload of 10mL/kg, and it may not be valid to extrapolate our results to patients who do not receive a similar fluid preload. The doses of ephedrine (6mg) and phenylephrine (100μg) used in our study were chosen empirically based on our clinical experience. After the start of our study, a potency ratio of phenylephrine to ephedrine was found to be 81.2 (95% CI 73.0-89.7).26 However, the vasopressors were administered as an infusion, which might not be applicable to bolus administration.

In conclusion, we have found that phenylephrine 100μg had similar vasopressor efficacy to ephedrine 6mg in the treatment of maternal hypotension during spinal anaesthesia for elective caesarean delivery. As in other studies, there were minor differences in fetal acid-base balance, but no difference in early neonatal neurobehavioural scores between the treatment groups.