Tan, Peng Chiong FRCOG; Khine, Pwint Phyu MBBS; Vallikkannu, Narayanan MOG; Omar, Siti Zawiah MOG
Hyperemesis gravidarum is practically defined as pernicious vomiting in pregnancy that requires hospitalization.1 It typically is associated with dehydration, electrolyte disturbances, and starvation with consequent weight loss. It affects 0.3% to 2.3% of the pregnant population,2 whereas nausea and vomiting of pregnancy is far more common, affecting up to 85% of pregnant women.3
The mainstays of therapy for hyperemesis gravidarum are rehydration, antiemetics, and enhancement of coping mechanisms.3 In their 2004 guidelines on nausea and vomiting in pregnancy, the American College of Obstetricians and Gynecologists recommends dimenhydrinate, metoclopramide, or promethazine intravenously as first-line antiemetic measures for women who have become dehydrated owing to hyperemesis gravidarum.3
In the context of management of hyperemesis gravidarum, promethazine has been compared with ondansetron4 and corticosteroids.5,6 Metoclopramide, in the context of the management of hyperemesis gravidarum has been described in the literature as being used in combination with diphenhydramine7 and promethazine,8 investigated for outpatient use as continuous subcutaneous infusion,9 and compared with hydrocortisone10 and with acupuncture.11
According to a recent Cochrane meta-analysis on interventions in early pregnancy, placebo-controlled trials have been reported for promethazine but not for metoclopramide.12 Although a comparative study of metoclopramide plus pyridoxine compared with promethazine has been reported in women with less severe nausea and vomiting of pregnancy, the addition of pyridoxine to only the metoclopramide arm did not permit a direct comparison with promethazine.13 Our objective was to compare promethazine with metoclopramide for hyperemesis gravidarum to determine whether metoclopramide may be a more effective antiemetic.13
The trial was conducted in a university hospital in Kuala Lumpur, Malaysia. About 210 new patients with hyperemesis gravidarum are admitted to our center per year.2 Ethical oversight was provided by the University of Malaya Medical Center Medical Ethics Committee (approval date November 21, 2007, reference number 625.10). This trial was performed in compliance with the Declaration of Helsinki. All participants provided informed written consent. The trial was conducted from November 25, 2008, to August 14, 2009.
Women hospitalized for the first time in their current pregnancies with presumed hyperemesis gravidarum were approached to participate in the trial by providers as soon as they were determined clinically to require intravenous antiemetic therapy. The decision to administer antiemetic therapy for presumed hyperemesis gravidarum could be made before investigations (eg, pelvic ultrasonography, urine culture) were completed as per our normal practice.14 In our center, as standard initial treatment, patients with hyperemesis gravidarum receive intravenous rehydration with saline (with the addition of potassium chloride as required if hypokalemic), oral thiamine 10 mg daily, and an intravenous antiemetic.14
Inclusion criteria were clinical hyperemesis gravidarum with dehydration and detectable ketonuria by urine dipstick at a gestation of 16 weeks or less. Exclusion criteria were multiple gestation, established nonviable pregnancy, preexisting medical condition that can cause nausea and vomiting (eg, culture-proven symptomatic urinary tract infection or dengue fever), gastrointestinal causes of vomiting (eg, gastroenteritis), medical causes of vomiting (eg, diabetic ketoacidosis), and known allergy to metoclopramide or promethazine.
A direct study comparing metoclopramide and promethazine was not available for sample-size calculation. We based our sample-size calculation on the hypothesis that metoclopramide will produce an overall well-being visual numerical rating scale score at 24 hours that is 1 unit better than that produced by promethazine (using a 10-point visual numerical rating scale), assuming a visual numerical rating scale standard deviation of 2, α=0.05, and 80% power. Sixty-four women were required in each arm. Taking into consideration that the distribution of the visual numerical rating scale score might not be normal and that the Mann-Whitney U test might have to be applied instead of the Student's t test, we factored in a 10% increase to the enrollment. Further, factoring in a 10% drop-out rate, we planned to enroll a total of 158 women to perform a suitably powered study.
Participants were recruited by providers as they were admitted to the gynecology ward. After obtaining consent, participants were assigned randomly by the sequential opening of numbered, sealed, opaque envelopes stating “Drug A” or “Drug B.” These numbered envelopes were prepared by an author (P.P.K.) in random blocks of four or eight using a computer-generated randomization sequence (performed at http://www.random.org).
Study drugs were placed in identical vials. Each vial contained a colorless 5-mL solution, and the vial was labeled as A or B to effect double blinding of drug allocation. The solutions contained either 10 mg of metoclopramide or 25 mg of promethazine. We swapped the contents of vials A and B between metoclopramide and promethazine periodically to strengthen blinding; that is, the study drugs were in vials labeled differently at different stages of the study. This was not revealed to providers.
Antiemetic therapy as randomly allocated was administered by slow injection into an indwelling intravenous catheter over 1 to 2 minutes by providers just after randomization and 8, 16, and 24 hours later for a full course of four doses as per trial protocol. Participants were instructed to chart vomiting episodes as they occurred during the 24-hour study period and to mark the nausea visual numerical rating scale (10 points, high score denoting more severe nausea) before initial administration of the study drug and 8, 16, and 24 hours after administration of the study drug. Also at 24 hours, participants were asked to mark their perceived well-being over the study period with a 10-point visual numerical rating scale (higher score=greater well-being) and to answer a preset questionnaire (yes or no answers) on symptoms experienced during the study period.
During the study period, the drug administered may have been switched from A to B or vice versa, without unblinding, if providers felt that it was clinically indicated. A switch would count as curtailment of treatment. At the conclusion of the 24-hour main study period, the allocated intravenous study antiemetic could be stopped or continued as allocated or open-label treatment (either intravenous or oral) instituted at the provider's discretion. Standard care of hyperemesis gravidarum patients otherwise was applied as previously described in our center.14–16
The preset primary outcomes were well-being visual numerical rating scale score and frequency of vomiting in the first 24 hours. Secondary outcomes included visual numerical rating scale scores for nausea at enrolment and at 8, 16, and 24 hours; adverse-symptoms profile; ketonuria status at the end of the 24-hour main study period; treatment curtailment during the main study period; total doses of intravenous antiemetic needed during hospitalization; admission-to-discharge interval; and time needed for intravenous rehydration.
Data were entered into SPSS 16 (SPSS Inc., Chicago, IL). Analysis was by intention to treat after exclusions for criteria infringements. Normal distribution of continuous data was checked with the one-sample Kolmogorov-Smirnov test. Normally distributed continuous data were analyzed with the Student's t test. Two-by-two categorical data sets were analyzed with the Fisher exact test and larger categorical data sets with the χ2 test; ordinal data and nonnormally distributed continuous data were analyzed with the Mann-Whitney U test. A repeated-measures analysis of variance was applied to the nausea visual numerical rating scale scores (four scores from each participant). Numbers needed to treat were generated using GraphPad QuickCalc (GraphPad Software Inc., La Jolla, CA). All tests were two-sided, and P<.05 was considered significant.
The recruitment flowchart of participants through the trial is shown in Figure 1. A total of 160 randomization envelopes were opened (including one envelope opened mistakenly, which was discarded unused; allocation to metoclopramide). Seventy-nine participants were assigned randomly to receive metoclopramide and 80 to receive promethazine. Ten women were excluded postrandomization, six from the metoclopramide arm and four from the promethazine arm. These exclusions were mostly due to study criteria infringements (shown in Fig. 1) that were ascertained only after randomization as a result of our pragmatic recruitment process. The 149 participants remaining started therapy as allocated.
Table 1 shows the characteristics of the participants in the two trial arms. Participants in the two arms were very similar. In addition to the characteristics shown in Table 1, all other parameters from the standard renal function test and full blood count were also similar. Random glucose and liver function tests and thyroid function test parameters where performed according to provider preference were also similar across both trial arms (data not shown).
Table 2 shows the outcomes analyses according to treatment allocation. Of the two preset primary outcomes over the 24-hour main study period, results were not significantly different: median vomiting frequency one (range 0–26) compared with two (range 0–26) (P=.81) and well-being visual numerical rating scale score 8 (range 1–10) compared with 7 (range 2–10) (P=.24) for the metoclopramide and promethazine arms, respectively. Of the secondary outcomes, the nausea visual numerical rating scale scores at 8, 16, and 24 hours postrandomization were very similar at each time point when assessed individually. A repeated-measures analysis of variance also showed no difference between the trial arms (P=.95) for nausea visual numerical rating scale scores. Length of hospital stay, persistence of ketonuria at 24 hours, overall treatment curtailment, duration of intravenous rehydration, and total doses of intravenous antiemetic given during hospitalization were not different between the trial arms. Of the yes or no responses to the questionnaire on symptoms at the end of the 24-hour main study period, drowsiness (58.6% compared with 83.6%, P=.001, number needed to treat to benefit [NNTb] 5), dizziness (34.3% compared with 71.2%, P<.001, NNTb 3), and dystonia (5.7% compared with 19.2%, P=.02, NNTb 8) were reported less frequently for metoclopramide than for promethazine. Difficulty in sleeping, dry mouth, diarrhea, headache, palpitations, and skin rash were reported in similar proportions across the trial arms. Thirteen women did not receive the full four doses of allocated study drug in the 24-hour main study period as per trial protocol (reasons for noncompliance are stated in the footnotes of Table 2). Of these women, 4 out of 73 (5.5%) compared with 9 out of 76 (11.8%) (P=.25) were assigned randomly to metoclopramide and promethazine, respectively, but metoclopramide was less likely to be associated with treatment curtailment owing to adverse events (0 of 73 [0%] compared with 7 of 76 [9.2%], P=.014).
Post hoc, taking the overall well-being (mean±standard deviation) visual numerical rating scale of 7.6±2.2 compared with 7.1±2.3 for metoclopramide compared with promethazine (Table 2), the calculated statistical power of our study was 38.5%. However, a small difference of 0.5 on a 10-point visual numerical rating scale is not likely to be clinically significant.
Intravenous 10-mg metoclopramide compared with intravenous 25-mg promethazine every 8 hours in the first 24 hours after hospitalization for hyperemesis gravidarum produced similar therapeutic effects. However, there were fewer reported side effects (eg, drowsiness, dizziness, and dystonia) among those randomly assigned to receive metoclopramide. Metoclopramide therapy was less likely to be curtailed owing to adverse events.
Although promethazine or metoclopramide intravenously are recommended as first-line antiemetic therapy for hyperemesis gravidarum,3 to our knowledge (PubMed search on September 23, 2009, using search terms “metoclopramide” and “promethazine” and “hyperemesis gravidarum” together without limitations), a trial directly comparing promethazine with metoclopramide for hyperemesis gravidarum has not been done. However, a three-armed trial of placebo compared with metoclopramide compared with promethazine (with the first dose of pethidine) for labor analgesia has shown that metoclopramide and promethazine are equally effective in reducing the incidence of nausea and vomiting but that the sedative effect was more persistent in the promethazine group.17 These findings are similar to our data.
Metoclopramide has a desirable effect on lower esophageal sphincter function, but promethazine is associated with evidence of increased gastroesophageal reflux.18 Our findings would indicate that esophageal reflux is probably not an important contributor to symptoms in hyperemesis gravidarum.
Our dosage for promethazine at 25 mg intravenously every 8 hours is within the recommended adult dose of up to 12.5 to 25 mg every 4 hours (Baxter Healthcare Corporation. Phenergan. Promethazine hydrochloride injection solution. Package insert). The mean weight of our study population was 54.3 kg (standard deviation ±9.3 kg). A placebo-controlled trial involving weight-adjusted metoclopramide and promethazine doses for postoperative nausea and vomiting in gynecologic surgery used 0.2 mg/kg for metoclopramide and 0.5 mg/kg for promethazine,19 which is very similar to our dosage. In a placebo-controlled trial of corticosteroids in hyperemesis gravidarum, participants received promethazine 25 mg and metoclopramide 10 mg intravenously every 6 hours for 24 hours as standard therapy,8 a treatment regime that is more intensive than our protocol. Post hoc analysis within our promethazine arm after stratifying participants into lower half compared with upper half of participants' body weight ranges showed no difference in reported drowsiness (83.8% compared with 84.4%, P=1.0), dizziness (73% compared with 71.9%, P=1.0), or dystonia (24.3% compared with 15.6%, P=.55). The same analysis also gave similar results for the metoclopramide arm. There was no indication that weight-adjusted dosing would change the adverse-effects profile.
There was more therapy curtailment because of adverse effects in the promethazine arm. Three of the seven episodes of treatment curtailment for adverse events were due to participants' refusal to accept a further injection because of pain during a previous injection. This occurred despite the fact that we used a concentration of 5 mg/mL of promethazine coupled to a slow injection rate more than 1 to 2 minutes within our protocol compared with the recommended maximum concentration of 25 mg per mL and an infusion rate of 25 mg per minute (Baxter Healthcare Corporation. Phenergan. Promethazine hydrochloride injection solution. Package insert). On September 16, 2009, the U.S. Food and Drug Administration warned of potential local tissue damage, including gangrene, from promethazine injection and mandated a black box warning on the drug insert.20
Our trial has limitations. The primary outcomes were over a 24-hour time scale only. Data from recent studies14–16 from our center indicate that between 24% and 32% of women with hyperemesis gravidarum were discharged after an overnight hospital stay, indicating that 24 hours could be appropriate to gauge antiemetic effectiveness in our patients. Furthermore, 76% of our study population needed only the four doses of intravenous antiemetic therapy during their hospitalizations. We recruited on the basis of presumed hyperemesis gravidarum before investigation results were available. Later discovery of trial criteria infringements resulted in postrandomization exclusion of 10 out of 159 (6.3%) women. However, findings were similar with these women included. There was a small number of participants whose data were incomplete. We looked at a broad range of adverse symptoms that could lead to a type 1 statistical error, but the pattern of our findings was consistent with the literature.17 It also should be noted that intravenous rehydration therapy alone might be effective in hyperemesis gravidarum, and a Cochrane review has indicated that no trial of treatments for hyperemesis gravidarum shows any evidence of benefit.12
Our patients with hyperemesis gravidarum21 had metabolic and biochemical characteristics comparable with other women in reported studies of hyperemesis gravidarum.22,23 Hence, our findings should be generalizable.
Intravenous metoclopramide or promethazine for hospitalized hyperemesis gravidarum patients produced similar therapeutic effects. Metoclopramide resulted in fewer reported side effects and treatment curtailment owing to adverse events. Intravenous metoclopramide is preferred over intravenous promethazine for the treatment of hyperemesis gravidarum.
1. Verberg MF, Gillott DJ, Al-Fardan N, Grudzinskas JG. Hyperemesis gravidarum, a literature review. Hum Reprod Update 2005;11:527–39.
2. Tan PC, Jacob R, Quek KF, Omar SZ. The fetal sex ratio and metabolic, biochemical, haematological and clinical indicators of severity of hyperemesis gravidarum. BJOG 2006;113:733–7.
3. Nausea and vomiting of pregnancy. Practice Bulletin No. 52. American College of Obstetricians and Gynecologists. Obstet Gynecol 2004;103:803–14.
4. Sullivan CA, Johnson CA, Roach H, Martin RW, Stewart DK, Morrison JC. A pilot study of intravenous ondansetron for hyperemesis gravidarum. Am J Obstet Gynecol 1996;174:1565–8.
5. Safari HR, Fassett MJ, Souter IC, Alsulyman OM, Goodwin TM. The efficacy of methylprednisolone in the treatment of hyperemesis gravidarum: a randomized, double-blind, controlled study. Am J Obstet Gynecol 1998;179:921–4.
6. Ziaei S, Hosseiney FS, Faghihzadeh S. The efficacy low dose of prednisolone in the treatment of hyperemesis gravidarum. Acta Obstet Gynecol Scand 2004;83:272–5.
7. Lacasse A, Lagoutte A, Ferreira E, Bérard A. Metoclopramide and diphenhydramine in the treatment of hyperemesis gravidarum: effectiveness and predictors of rehospitalisation. Eur J Obstet Gynecol Reprod Biol 2009;143:43–9.
8. Yost NP, McIntire DD, Wians FH Jr, Ramin SM, Balko JA, Leveno KJ. A randomized, placebo-controlled trial of corticosteroids for hyperemesis due to pregnancy. Obstet Gynecol 2003;102:1250–4.
9. Buttino L Jr, Coleman SK, Bergauer NK, Gambon C, Stanziano GJ. Home subcutaneous metoclopramide therapy for hyperemesis gravidarum. J Perinatol 2000;20:359–62.
10. Bondok RS, El Sharnouby NM, Eid HE, Abd Elmaksoud AM. Pulsed steroid therapy is an effective treatment for intractable hyperemesis gravidarum. Crit Care Med 2006;34:2781–3.
11. Neri I, Allais G, Schiapparelli P, Blasi I, Benedetto C, Facchinetti F. Acupuncture versus pharmacological approach to reduce Hyperemesis gravidarum discomfort. Minerva Ginecol 2005;57:471–5.
12. Jewell D, Young G. Interventions for nausea and vomiting in early pregnancy. The Cochrane Database of Systematic Reviews 2003, Issue 4. Art. No.: CD000145. DOI: 10.1002/14651858.CD000145.
13. Bsat FA, Hoffman DE, Seubert DE. Comparison of three outpatient regimens in the management of nausea and vomiting in pregnancy. J Perinatol 2003;23:531–5.
14. Tan PC, Yow CM, Omar SZ. A placebo-controlled trial of oral pyridoxine in hyperemesis gravidarum. Gynecol Obstet Invest 2009;67:151–7.
15. Tan PC, Jacob R, Quek KF, Omar SZ. Indicators of prolonged hospital stay in hyperemesis gravidarum. Int J Gynaecol Obstet 2006;93:246–7.
16. Tan PC, Tan NC, Omar SZ. Effect of high levels of human chorionic gonadotropin and estradiol (E2) on the severity of hyperemesis gravidarum. Clin Chem Lab Med 2009;47:165–71.
17. Vella L, Francis D, Houlton P, Reynolds F. Comparison of the antiemetics metoclopramide and promethazine in labour. Br Med J (Clin Res Ed) 1985;290:1173–5.
18. Brock-Utne JG, Rubin J, Welman S, Dimopoulos GE, Moshal MG, Downing JW. The action of commonly used antiemetics on the lower oesophageal sphincter. Br J Anaesth 1978;50:295–8.
19. Sandhya, Yaddanapudi LN. Evaluation of two antiemetic agents during outpatient gynaecological surgery. Singapore Med J 1994;35:271–3.
21. Tan PC, Jacob R, Quek KF, Omar SZ. Readmission risk and metabolic, biochemical, haematological and clinical indicators of severity in hyperemesis gravidarum. Aust N Z J Obstet Gynaecol 2006;46:446–50.
22. Goodwin TM, Montoro M, Mestman JH. Transient hyperthyroidism and hyperemesis gravidarum: clinical aspects. Am J Obstet Gynecol 1992;167:648–52.
23. Bashiri A, Neumann L, Maymon E, Katz M. Hyperemesis gravidarum: epidemiologic features, complications and outcome. Eur J Obstet Gynecol Reprod Biol 1995;63:135–8.