Reduction of severity of pruritus after elective caesarean section under spinal anaesthesia with subarachnoid morphine: a randomised comparison of prophylactic granisetron and ondansetron

Introduction 

Subarachnoid morphine is highly effective in the management of acute pain after caesarean delivery under neuraxial anaesthesia,1 but the incidence of pruritus has been reported to be between 60-100%2 and is a common cause of maternal dissatisfaction. The exact aetiology of pruritus is unclear. Modulation of serotoninergic pathways is one possible mechanism and 5-hydroxytryptamine type 3 (5-HT3) antagonists have been used successfully to treat opioid-induced pruritus.3, 4, 5 Results conflict whether prophylactic ondansetron, a 5-HT3 antagonist, reduces the incidence of subarachnoid opioid-induced pruritus in patients undergoing caesarean delivery under spinal anaesthesia.6, 7 A possible reason for poor prophylactic efficacy of ondansetron could be that the effect of morphine-induced pruritus peaks at 6-9h, whereas the duration of action of ondansetron is 8h. Granisetron, a highly selective 5-HT3 receptor antagonist, has a longer duration of action than ondansetron and may be more effective in reducing pruritus. However, the overall incidence of pruritus has not been shown to be reduced with granisetron in patients undergoing caesarean section who received subarachnoid morphine.9 Although the overall incidence of pruritus is not decreased, what is not clear is whether the prophylactic administration of a long acting 5-HT3 antagonist like granisetron reduces the severity of pruritus, thus improving maternal satisfaction. To our knowledge, no studies have been powered to detect a reduction in severity of pruritus with a long acting 5-HT3 antagonist.

The objective of this randomized double-blind trial was to determine if prophylactic administration of ondansetron 8mg or granisetron 3mg reduced the severity of pruritus after elective caesarean section under spinal anaesthesia. The secondary objectives were to determine if request for rescue anti-pruritic agents, nausea and vomiting was reduced and if maternal satisfaction was improved.

Methods 

After obtaining institutional ethics committee approval and written informed consent, 80 ASA I-II parturients with a singleton pregnancy, undergoing elective caesarean delivery under spinal anaesthesia were recruited. Patients with preeclampsia, major systemic disease, dermatological conditions, pre-existing pruritus, known allergy to any of the medications used or any condition that contraindicated spinal anaesthesia were excluded from the study.

Patients were randomly allocated using computer generated random numbers to one of two groups: group O (ondansetron 8mg i.v.) and group G (granisetron 3mg i.v.). The study drug was administered after delivery of the baby following clamping of the umbilical cord.

According to routine practice in our hospital, ranitidine 150mg orally was given on the morning of the caesarean section, followed by 0.3% sodium citrate 30mL 30min before surgery. Spinal anaesthesia was performed at the L3-4 interspace with a 25-gauge pencil-point (Whitacre) needle and 0.5% hyperbaric bupivacaine 10mg with fentanyl 25μg plus preservative free morphine 150μg was administered. The patient was positioned supine with left uterine displacement using a wedge under the right hip. Surgery was allowed to proceed once bilateral sensory block to pin prick to T5 was achieved. Monitoring included continuous electrocardiogram, pulse oximetry and non-invasive blood pressure. After the baby was delivered and the umbilical cord clamped, oxytocin 5 units was given as a slow i.v. bolus. Cefuroxime 1g i.v. was administered as antimicrobial prophylaxis. Diclofenac 100mg and paracetamol 1g were administered rectally at the end of the case. Postoperatively, all patients received diclofenac 100mg every 16h rectally, paracetamol 1g with codeine 60mg orally every 6h for pain relief. Postoperative fasting continued for 6h after surgery. All patients received i.v. Hartmann’s solution at 80mL/h until oral fluids were resumed.

Patients were evaluated by an investigator blinded to treatment group at 1, 2, 4, 6, 8 and 24h after administration of spinal anaesthesia. Visual analogue scales (VAS) (0: no symptoms, 10cm: worst possible) were used to assess pruritus and pain. Nausea and vomiting was graded on a 4 point scale (1: no symptoms, 2: nausea, 3: nausea with vomiting relieved by treatment, 4: nausea and vomiting unrelieved by treatment). Patients were asked after 24h to rate their satisfaction with pruritus on a four-point scale (very dissatisfied, dissatisfied, satisfied or very satisfied). A response of either satisfied or very satisfied was considered as successful anti-pruritic prophylaxis. Adverse reactions to ondansetron and granisetron (headaches, palpitations, extra-pyramidal signs) were recorded. Patients with a pruritus VAS >6 were treated with chlorpheniramine 4mg orally and, if the pruritus was not relieved, with incremental doses of naloxone 0.02mg i.v. Nausea and vomiting were treated with prochlorperazine 12.5mg i.m. 8-hourly. Patients with breakthrough pain were treated with oxycodone 10mg orally.

Results 

A total of 80 patients were studied with 40 patients recruited to each group. There was no protocol violation and all the recruited patients were included in the analysis. There were no differences in patient characteristics or preoperative data between the groups (Table 1).

Table 1. Patient characteristics

		Ondansetron	Granisetron	P
		(n=40)	(n=40)
	Age (years)	33.4 (0.76)	33.1 (0.79)	0.79
	Weight (kg)	77.7 (2.73)	79.2 (2.38)	0.67
	Height (m)	1.63 (0.01)	1.63 (0.09)	0.99
	Gestational age (weeks)	38.9 (0.25)	39.1 (0.13)	0.44

Data are mean (SD).

There was no difference in the overall incidence of pruritus between the groups (group G 92.5% vs. group O 95%). The mean VAS for pruritus over 24h after caesarean section is shown in Fig. 1. The mean VAS for pruritus at 8h was 1.64 in group G, and 3.41 in group O, mean difference 1.77 (95% C.I. 0.80 to 2.70; P=0.01). This difference was also seen at 24h where the mean VAS of pruritus in group G was 0.51 compared to 1.41 in group O, mean difference 0.9 (95% C.I. 0.22 to 1.59; P=0.01). The use of rescue anti-pruritic agents was significantly greater in group O than in group G only within the time frame of 8 to 24h (P=0.03). The need for rescue anti-pruritic agents and the time intervals during which they were requested are presented in Table 2. No patient required intravenous naloxone. The incidence of nausea and vomiting, and use of anti-emetic agents are shown in Table 3. Except at 4h where group G had significantly less nausea, there was no difference between the groups. Pain scores (Fig. 2) and the need for rescue analgesia (8 patients in group G compared with 10 in group O) were similar in the two groups (P=0.79).

View Large Image Download to PowerPoint

Fig. 1. Mean VAS score for pruritus over time. Vertical bars represent 95% C.I. ∗P<0.05 refers to between groups comparisons.

Table 2. The need for rescue anti-pruritic agents

		Granisetron	Ondansetron	P
		n=40	n=40
	Use of oral chlorphenarimine
	0 to 8 hours	8 (20%)	13 (32.5%)	0.31
	8 to 24 hours	0 (0%)	6 (15%)	0.03⁎
	Use of intravenous naloxone
	0 to 8 hours	0	0	–
	8 to 24 hours	0	0	–

Significantly more requests for rescue anti-pruritic agents in the ondansetron group.

⁎ P<0.05.

Table 3. Incidence of nausea and vomiting and the use of rescue anti-emetics

			Granisetron	Ondansetron	P
			n=40	n=40
	Patients experiencing nausea and vomiting		12 (30%)	20 (50%)	0.11
	Patients requiring rescue anti-emetics		6 (15%)	11 (27.5%)	0.27
	Nausea score	1h	1 [1-1]	1 [1-1]	0.33
		2h	1 [1-1]	1 [1-1]	0.40
		4h	1 [1-1]	1 [1-2]	0.04
		6h	1 [1-1]	1 [1-2]	0.49
		8h	1 [1-1]	1 [1-1]	0.71
		24h	1 [1-1]	1 [1-1]	0.99

Data are n (%) or median [inter-quartile range].

View Large Image Download to PowerPoint

Fig. 2. Mean VAS score for pain over time. There are no significant differences between groups. Vertical bars represent 95% C.I.

Satisfaction scores 24h after surgery are presented in Table 4 and Fig. 3. Significantly more patients scored satisfied or very satisfied in group G than in group O (P=0.02).

Table 4. Satisfaction scores

		Granisetron	Ondansetron	P
		(n=40)	(n=40)
	Patients scoring dissatisfied or very dissatisfied	2 (5%)	10 (25%)	0.02
	Patients scoring satisfied or very satisfied	38 (95%)	30 (75%)	NS

Data are n (%).

View Large Image Download to PowerPoint

Fig. 3. Patient’s response to the severity of pruritus experienced. ∗Significantly more patients were dissatisfied in group O (P<0.05).

No adverse effects such as headache, palpitations or extrapyramidal signs were observed following administration of either ondansetron or granisetron.

Discussion 

We found prophylactic administration of granisetron to be effective in reducing pruritus after elective caesarean section in women who received subarachnoid morphine. There was less demand for rescue anti-pruritic agents in patients who received granisetron compared with ondansetron. Furthermore, patients in the granisetron group had better satisfaction scores. Although the severity of pruritus was reduced, there was no difference in the incidence of pruritus between the groups. Except for the difference at 4h, which we did not consider as clinically significant, there was no difference in the incidence of nausea and vomiting in the granisetron group compared to the ondansetron group.

Elective caesarean section is commonly performed with neuraxial anaesthesia by administering a combination of hyperbaric bupivacaine, fentanyl and a longer acting opioid, either morphine or diamorphine into the subarachnoid space. This provides excellent conditions for surgery and effective postoperative analgesia. Unfortunately the incidence of pruritus in the obstetric population, especially after subarachnoid opioids, can be very high ranging between 70 and 93%.4, 5, 6, 7, 9 This is consistent with our population among whom the incidence of pruritus in ondansetron and granisetron groups was 92.5% and 95% respectively. One of the possible mechanisms of pruritus is related to direct stimulation of 5-HT3 receptors by morphine in the dorsal horn of the spinal cord and medulla, therefore 5-HT3 antagonists may be beneficial in its treatment.10 Previous studies by Yeh et al.4 and Charuluxananan et al.5 demonstrated that prophylactic ondansetron reduced the frequency of pruritus in patients undergoing caesarean section. Conversely, three recent studies6, 7, 9 comparing longer acting 5-HT3 antagonists reported the failure of ondansetron, tropisetron and granisetron to prevent pruritus after caesarean section. These conflicting results may be due to the larger doses of subarachnoid morphine administered and to different scales and definitions used to assess pruritus. More importantly, these previous studies were powered to detect a difference in incidence and not reduction in severity of pruritus.

Another reason for the lack of efficacy of prophylactic ondansetron shown in previous studies may be its relatively short duration of action of approximately 8h. This does not coincide with the peak incidence of morphine-induced pruritus at 6-9h after caesarean section,8 thereby rendering it ineffective as a prophylactic drug. Our study showed that granisetron, a highly specific 5-HT3 antagonist with a duration of action of 24h was more effective in reducing the severity of pruritus between 8 and 24h.

Although the severity of pruritus was reduced by granisetron, there was no difference in the overall incidence of pruritus in the two study groups. It is postulated that there may be several mechanisms causing pruritus after caesarean delivery such as the interaction of oestrogen with opioid receptors in the spinal cord, activation of nucleus trigeminalis, and the antagonism of inhibitory neurotransmitters.11 As multiple mechanisms are involved, it is unlikely that a therapeutic agent with a presumed single mode of action such as granisetron can be totally effective. What we have demonstrated is that, although pruritus may not be completely eliminated, it can be reduced to a more tolerable degree thus improving patient satisfaction.

Unlike previous studies, we did not exclude women who were breast-feeding as there are no known reports of adverse neonatal effects as a result of transfer through breast milk. There are no human data on the excretion of ondansetron or granisetron to breast milk or neonatal transfer, but it is unlikely that breast feeding carries a significant risk.12 Furthermore, both drugs were administered as a single dose after clamping of the umbilical cord. We anticipate that the excretion into colostrum and eventual transfer to the neonate to be clinically insignificant.

A shortcoming of our study is that a placebo group was not included to evaluate the efficacy of ondansetron and granisetron in reducing pruritus. However, our local ethics committee deemed it unethical to withhold anti-emetic agents in peripartum patients who have received subarachnoid morphine. A recent systematic review by Bonnet et al. published after completion of our trial concluded that 5-HT3 receptor antagonists may be effective in preventing neuraxial opioid-induced pruritus, but that more trials were needed to confirm this finding as there seemed to be a publication bias in the current literature.13 The inclusion of a placebo group would have addressed this issue. Another possible confounding factor in our study was the routine administration of diclofenac to all patients, as non steroidal anti-inflammatory drugs may be effective in reducing pruritus.14 Because of the well proven therapeutic efficacy of diclofenac, it was also deemed unethical to withhold the drug for the purpose of this study.

It would have been interesting to compare the pruritus VAS scores between 8 and 24h. However, the 12 and 16h time points frequently fell after 22.00h and before 07.00h, and we did not have sufficient manpower or resources to ensure that data were reliably collected at these times. Our data showed that the pruritus VAS score was already significantly different at 8h, persisting up to 24h. We therefore assumed that the VAS score would be continuously lower between 8 and 24h.

In conclusion, we have shown that granisetron reduces the severity of pruritus after subarachnoid opioids compared to ondansetron. To our knowledge, there is no other study that has been powered to detect a decrease in severity of pruritus with the use of long acting 5-HT3 antagonists. We recommend that granisetron be administered prophylactically to reduce pruritus and improve satisfaction scores after caesarean section where subarachnoid morphine 150μg (plus rectal diclofenac 100mg) is administered for postoperative analgesia. Further research is needed to elucidate the mechanisms, mediators, neural pathways and pathophysiology of this symptom. The optimal dose of subarachnoid opioid that provides maximal pain relief during and after caesarean section with minimal side effects should also be determined. Research is also needed to assess if 5-HT3 antagonists are of benefit with other opioids such as subarachnoid diamorphine, which is more commonly used in the United Kingdom.