| The benefit of steroid treatment in women with severe hypaemesis gravidarum, in terms of reduction in vomiting and need for intravenous rehydration is unproven. |
Citation/s:
Nelson-Piercy C, Fayers P, De Swiet M. Randomised, double-blind placebo controlled trial of corticosteroids for the treatment of hyperemesis gravidarum. Br J Obstet Gynaecol 2001, 108, 9-15.
Lead author's name and fax: Nelson-Piercy C
Three-part Clinical Question: Do steroids reduce the frequency the vomiting and the need for intravenous fluid replacement in women with severe hyperemesis gravidarum?
Search Terms: (hyperemesis gravidarum (MeSH) AND steroids (MeSH)) and ([double (tw) and blind*(tw)] OR placebo(tw))
The Study:
Double-blinded concealed randomised controlled trial with intention-to-treat.
The Study Patients: Pregnant women with severe hyperemesis - n&v < 12 weeks gestation, dependent on iv fluids >/= 1 week (1st admission) or >24 hours (subsequent admission), regular treatment with an antiemetic, ketonuria on admission, a negative MSU and normal random blood glucose. Vomiting >/= x2/day and on thiamine supplementation. No Rx with steroids < 2 months of study or CI to steroids and not clinically thyrotoxic.
Control group (N = 12; 12 analysed): Placebo tablets b.d.
(if still vomiting >72 hours then converted to i.v. placebo (normal saline)
Experimental group (N = 13; 12 analysed): Prednisolone 20 mg orally b.d.
(if still vomiting >72 hours than converted to iv hydrocortisone 100mg b.d.)
The Evidence:
| Outcome | Time to Outcome | CER | EER | RRR | ARR | NNT |
| Vomiting | 1 week | 0.583 | 0.417 | 28% | 0.166 | 6 |
|---|---|---|---|---|---|---|
| 95% Confidence Intervals: | -39% to 96% | -0.229 to 0.561 | NNT = 2 to INF; NNH = 4 to INF | |||
| Severe vomiting (>/= x5 times/day) | 1 week | 0.417 | 0.167 | 60% | 0.250 | 4 |
| 95% Confidence Intervals: | -24% to 100% | -0.100 to 0.600 | NNT = 2 to INF; NNH = 10 to INF | |||
| Dependent on i.v. fluids | 1 week | 0.250 | 0.250 | 0% | 0.000 | INF |
| 95% Confidence Intervals: | -100% to 100% | -0.346 to 0.346 | NNT = 3 to INF; NNH = 3 to INF | |||
Comments:
Unrealistic original sample size calculation (sample size 45) and poor recruitment. Study therefore well underpowered and consequently minimal benefit in terms of informing the clinician. Short duration of follow up (1 week).
Appraised by: Dr A Musa and Dr TJ Clark; 26 April 2002
Email: musa_80@hotmail.com
Kill or Update By: 2005