Published on November 25th, 2012 | by Daniel Jolley3
Metoclopramide: it actually works!?
Metoclopramide had long been written off by many anesthetists and anesthesiologists, aware of trials and meta-analyses that show no or limited effect in treating or preventing nausea and vomiting — in particular limited ability to prevent post-operative nausea and vomiting (PONV). Most recently Henzi, Walder and Tramèr (1999) were able to show only very limited benefit for metoclopramide 10 mg in preventing vomiting and no significant effect in preventing nausea in adults. 1, 2
What is metoclopramide?
Metoclopromaide is a benzamide, predominately used for antiemesis and its gastric prokinetic effect. It is marketed under the names Maxalon®, Pramin® and Reglan® in various countries. Although considered an old drug its antiemetic action was first identified in 1964 by French doctors Justin-Besançon and Laville. 3 (In contrast the analgesic tramadol is often considered a ‘new’ drug outside of Europe, but was launched by Grünenthal GmbH in 1977.)
Metoclopramide readily crosses the blood-brain barrier where it mediates anti-emetic effects primarily as a dopamine D2 antagonist in the chemoreceptor trigger zone (CTZ — located in the area postrema of the 4th ventricle). Metoclopramide also has mixed 5-HT3 receptor antagonist and 5-HT4 receptor agonist actions. The former may contribute to anti-emesis at higher doses and the later to its pro-kinetic effects. Muscarinic cholinergic actions have also been identified, both through increasing acetylcholine release and by increasing receptor sensitivity to acetylcholine in the upper GI tract — further contributing to the pro-kinetic effect.
The rise of the 5HT3 antagonist in our PONV arsenal has further pushed metoclopramide down a growing list of options. The Society of Ambulatory Anesthesia no longer even recommends the use of metoclopramide for PONV prophylaxis:
“Some therapies have proven ineffective for PONV prophylaxis. These include metoclopramide when used in standard clinical doses (10 mg IV)…” 4
Metoclopramide’s remaining virtues appeared limited to speeding gastric emptying 5 and it’s incredibly low cost — in many health systems cheaper in fact than an ampoule of saline.
Enter Fujii’s fraud
Kranke, Apfel and team were the first to highlight that a large body of research contributed by a single researcher and institution could significantly shift the findings of a meta-analysis. 6 The dominant institution that worried Kranke et al. was Toride Kyodo General Hospital in Japan — and the researcher was Yoshitaka Fujii.
Fast forward eleven years, and while the Fujii fraud is now well known (see “Fujii, anesthesia and research fraud”) the impact of Fujii’s research on assumed anesthesia truths is still emerging. The impotence of metoclopramide is the first accepted truth to be revisited.
Re-analysing without Fujii’s studies
De Oliveira and team performed a systematic review of 30 studies and over 3000 patients examining metoclopramide 10 mg, specifically excluding those with “questioned validity” from Fujii. 7 They showed metoclopramide to be significantly better than previously thought, effective in reducing both the incidence of PONV, and even nausea and vomiting when examined individually.
Metoclopramide reduced the incidence of 24 h PONV compared with control, odds ratio (OR) [95% confidence interval (CI)] of 0.58 (0.43-0.78), number needed to treat (NNT)=7.8. When evaluated as separate outcomes, metoclopramide also decreased the incidence of nausea over 24 h, OR (95% CI) of 0.51 (0.38-0.68), NNT=7.1, and vomiting over 24 h, OR (95% CI) of 0.51 (0.40-0.66), NNT=8.3.
De Oliveira demonstrated a significant effect in preventing early PONV (NNT 7.6) — in particular for preventing nausea alone (NNT 5.9) — though less effect in preventing early vomiting (NNT 10.5).
While these results still fall a little short of the performance of 5HT3-antagonists (ondansetron reportedly has a NNT=6 for nausea prophylaxis and NNT=7 for vomiting) the small difference in efficacy may be made up for by the dramatic lesser cost. 8 De Oliveria’s results for metoclopramide compare favourably with 8-10 mg of dexamethasone, having a reported NNT of 7.1 but a terribly wide confidence interval (95% CI 4.5–18).
Clinicians have been reluctant to use metoclopramide not just for its perceived lack of efficacy but also believing that it carries a higher risk of side effects than other antiemetics. However De Oliveira’s review did not detect any increase in minor side effects such as headache, dizziness or sedation when compared with saline controls, and showed no statistically significant increase in extrapyramidal reactions. 9
As with all meta analyses, conclusions must be tempered with an appreciation of the limitations when grouping different studies together. Although this review includes studies involving a range of different surgeries the authors show a low measured heterogeneity in the results — giving us some confidence in the generalizability of metoclopramide prophylaxis to different types of surgery and anesthesia techniques.
Quantifying Fujii’s effect
John Carsisle followed up his impressive exposure of the Fujii fraud 10 with a lap of honor quantifying the differences between Fujii’s anti-emetic research and results from other authors.
Compared to other authors Fujii showed 5HT3-antagonists granisetron and ramosetron to be more effective 11 and droperidol to be less effective for PONV prophylaxis. Fujii also showed rescue anti-emesis to be less frequently required after granisetron than for droperidol and metoclopramide, whereas other researchers showed no difference. Finally and of most interest, Fujii’s papers suggested a synergistic action between anti-emetics absent from the research of others — in fact other research suggests that there may be antagonism between certain anti-emetic drugs. 12
“…if we exclude the data of Fujii et al. … Granisetron remains more effective than metoclopramide for all four outcomes, with significant differences vs droperidol persisting for two outcomes (vomiting, rescue) but disappearing for nausea and the combined ‘nausea and vomiting’ outcomes. Against placebo, the effects of granisetron and ramosetron are less … Postoperative nausea and vomiting after granisetron triggers rescue antiemesis at the same rate as droperidol and metoclopramide, instead of less often. The unique characteristic of granisetron to act synergistically with other antiemetics disappears.”
Metoclopramide & cesarean section
If we take a few steps closer in the ‘metoclopramide question’ a specific surgical procedure offers further insight: cesarean section under regional anesthesia. This group of patients suffers both a high incidence of PONV and uniquely also a high incidence of intra-operative nausea and vomiting (IONV).
Mishriky and Habib reviewed this very question in a meta-analysis, concluding that metoclopramide was an effective agent for both IONV & PONV prophylaxis during cesarean section — a finding that probably does not surprise the older and more experienced obstetric anesthesiologists among us. 13
They showed 10 mg of metoclopramide to be effective at reducing intra-operative nausea and vomiting whether given before block placement or after delivery — although the effect was greater if metoclopramide was given before the block (RR 0.27 vs 0.38, & 0.14 vs 0.34 for ION and IOV respectively) 14 Metoclopramide was also associated with a lower incidence of early PONV (0-3 h) and PONV over all (0-24 h & 3-24 h). Reassuringly, extra-pyramidal reactions were not reported in any of the 11 studies totally 702 patients.
Though we need to bear in mind that there was a lot of heterogeneity between the included studies: spinal (9 studies) vs epidural (2); before block (4) or after delivery (7) timing; and most notably presence (4) of absence (7) of neuraxial opioids. This diversity resulted from the inclusion of many older studies — unfortunately metoclopramide has not been an exciting research topic for some time!
Additionally, two studies from Fujii were unfortunately included in the analysis, 15 both investigating anti-emetics given after delivery while under spinal anesthesia. Though sub-groups that did not include these studies (pre-block metoclopramide) showed even greater beneficial effect for metoclopramide than the sub-group (metoclopramide after delivery) that did include them.
Even in the face of the broad study diversity, sub-group analyses showed metoclopramide resulted in significant reductions in intra-operative nausea and vomiting with quite low NNTs — for example, pre-spinal metoclopramide investigated in three studies resulted in a NNT of 2 and 5 for avoiding ION and IOV respectively. 16 17
Finally, the effectiveness of metoclopramide for IONV and PONV prophylaxis during cesarean sections was borne out by a recent Cochrane review from Griffiths et al. 18 While the reviewers noted the superiority of 5HT3-antagonists, the efficacy of metoclopramide for IONV prophylaxis was again demonstrated. Importantly Fujii’s studies were not included in this analysis.
An old drug with old tricks?
Modern 5HT3-antagonist anti-emetics are still superior to metoclopramide: they are both more efficacious and have better side effect profiles. But for this benefit there is a substantial cost that must be considered when choosing an anti-emetic — these drugs are around 20 times more expensive than metoclopramide in most markets. ‘Number needed to treat’ when applied to large at-risk groups must also be balanced with our capacity to treat in a medical world with finite health budgets.
Metoclopramide — worthwhile after all?
- Henzi I, Walder B, Tramèr MR. Metoclopramide in the prevention of postoperative nausea and vomiting: a quantitative systematic review of randomized, placebo-controlled studies. Br J Anaesth. 1999 Nov;83(5):761-71. ↩
- It’s also interesting to note that there was no dose responsive effect regardless of route. NNT to prevent early (<6h) and late (<48h) vomiting were 9.1 (95% CI 5.5-27) and 10 (6-41) respectively. In children the best documented regimen was 0.25 mg/kg. NNT to prevent early vomiting was 5.8 (3.9-11); there was no effect on late vomiting. There was only a single documented case of extrapyramidal side effects out of 3260 patients, giving an incidence of 0.03%. ↩
- Justin-Besançon L, Laville C.C R Seances Soc Biol Fil. 1964;158:723-7. ↩
- Gan TJ, et al. Society for Ambulatory Anesthesia guidelines for the management of postoperative nausea and vomiting. Anesth Analg. 2007 Dec;105(6):1615-28 ↩
- Teramoto H, et al. Assessment of gastric emptying and duodenal motility upon ingestion of a liquid meal using rapid magnetic resonance imaging. Exp Physiol. 2012 Apr;97(4):516-24. ↩
- Kranke, et al. showed data from Fujii’s studies suggested granisetron was both more effective than other studies and exhibited dose-related effects — which then significantly effected the results of the meta-analysis. Kranke P, Apfel CC, Eberhart LH, Georgieff M, Roewer N. The influence of a dominating centre on a quantitative systematic review of granisetron for preventing postoperative nausea and vomiting. Acta Anaesthesiol Scand. 2001 Jul;45(6):659-70 ↩
- De Oliveira GS, Castro-Alves LJ, Chang R, Yaghmour E, McCarthy RJ Systemic metoclopramide to prevent postoperative nausea and vomiting: a meta-analysis without Fujii’s studies. Br J Anaesth. 2012 Sep 25; ↩
- In most markets metoclopramide is around 20 times less expensive than 5HT3-antagonists. ↩
- It should be noted though that with N=3328 the study was under-powered to detect or accurately report on the incidence of dystonic and extrapyramidal reactions. ↩
- Carlisle JB. The analysis of 168 randomised controlled trials to test data integrity. Anaesthesia. 2012 May;67(5):521-37. ↩
- Fujii, the publisher of the bulk of granisetron RCT, claimed granisetron was 1.3 to 2.5 times more effective than others. ↩
- Carsisle writes: “…synergism between antiemetics was found only in RCTs by Fujii et al. (for both droperidol and granisetron), whereas antagonism with other antiemetics was found in RCTs by others for dexamethasone, droperidol and ondansetron.” ↩
- Mishriky BM, Habib AS. Metoclopramide for nausea and vomiting prophylaxis during and after Caesarean delivery: a systematic review and meta-analysis. Br J Anaesth. 2012 Mar;108(3):374-83. ↩
- ”Administration of metoclopramide (10 mg) resulted in a significant reduction in the incidence of ION and IOV when given before block placement [relative risk (RR) (95% confidence interval, 95% CI)=0.27 (0.16, 0.45) and 0.14 (0.03, 0.56), respectively] or after delivery [RR (95% CI)=0.38 (0.20, 0.75) and 0.34 (0.18, 0.66), respectively]. The incidence of early (0-3 or 0-4 h) PON and POV [RR (95% CI)=0.47 (0.26, 0.87) and 0.45 (0.21, 0.93), respectively] and overall (0-24 or 3-24 h) PON (RR 0.69; 95% CI 0.52, 0.92) were also reduced with metoclopramide. “ ↩
- Unfortunately Fujji’s fraud had not been identified at the time of this systematic review’s publication. ↩
- Three studies administered metoclopramide before block placement and reported on ION and IOV. All three studies used spinal anaesthesia. Combining results from the three studies showed that metoclopramide caused a statistically significant reduction in the incidence of ION (RR 0.27; 95% CI 0.16, 0.45) and IOV (RR 0.14; 95% CI 0.03, 0.56) when compared with placebo. The NNT for ION and IOV was 2 and 5, respectively. … Six studies gave metoclopramide after delivery and reported on ION and IOV. Four studies used spinal anaesthesia, while two used epidural anaesthesia. Pooled results from the six studies showed a statistically significant reduction in the incidence of post-delivery ION (RR 0.38; 95% CI 0.20, 0.75) and IOV (RR 0.34; 95% CI 0.18, 0.66) in the metoclopramide group. The NNT for ION and IOV was 4 and 7, respectively. ↩
- The only sub-group which did not show some benefit was the single study investigating intrathecal morphine — there was, perhaps unsurprisingly, no reduction in PONV. ↩
- Griffiths JD, Gyte GM, Paranjothy S, Brown HC, Broughton HK, Thomas J. Interventions for preventing nausea and vomiting in women undergoing regional anaesthesia for caesarean section Cochrane Database Syst Rev. 2012 Sep 12;9:CD007579. (full text) ↩