New research should be presented in the context of existing literature: the case of misoprostol third stage of labour trials

RHL Editorial by Gülmezoglu AM, Villar J, Khanna J, Schulz KF, Lumbiganon P, Mittal S, Hofmeyr GJ, Cheng L

UNDP/UNFPA/WHO/World Bank Special Programme of Research, Development and Research Training in Human Reproduction, WHO, Geneva, Switzerland

In reporting new trial findings not all investigators follow the required practice of discussing their new results in the context of results of systematic reviews, where they exist. Systematic reviews present overviews of all relevant randomized controlled trials and use meta-analysis to increase the precision of individual trial estimates (i.e. narrower 95% confidence intervals). The greater precision allows for more accurate conclusions to be drawn. For example, in individual trials the difference between two treatments may be found to be statistically non-significant, but in a systematic review, pooled data from several trials may show a similar point estimate but the difference could be statistically significant. Discussing results of new trials in the context of a systematic review is especially important for trials designed to test equivalence of treatments. Failure to do so may help to perpetuate erroneous conclusions, possibly resulting in adverse consequences for practice. The case of misoprostol for the prevention of postpartum haemorrhage (PPH) described below illustrates this problem.

Over the last decade, misoprostol has arguably been one of the most researched and debated drugs in the field of reproductive health. Its potent cervical priming and uterotonic effects have been found to be effective in medical abortion and labour induction at term. Its use in the prevention of PPH, on the other hand, is somewhat controversial. In 2001, WHO conducted a large multicentre equivalence trial to evaluate the benefits and harms of misoprostol compared with oxytocin 10 IU for the prevention of PPH (1). Prior to this trial, in 1997, WHO had published a systematic review of all prostaglandin trials of PPH prevention (2); this review was updated in 2002 to reflect the findings of the new WHO trial (3). In addition, in order to ensure the widest possible availability of the findings of the Cochrane systematic review, WHO also published the same review in a print journal (4).

The main question for trials of misoprostol for PPH prevention has been whether misoprostol is as effective as the standard injectable uterotonics, such as oxytocin. Given misoprostol's advantages - oral administration, low cost, long shelf-life, etc. - demonstration of its non-inferiority to standard injectable uterotonics would give it a clear advantage over its rivals.

Equivalence trials typically require larger sample sizes. Results of underpowered equivalence trials are more likely to be interpreted as “equivalent treatments” because the difference between the interventions being compared does not reach statistically significant levels of conventional P <0.05 simply because of the small sample size. The remedies against this fallacy are to conduct adequately powered trials or to combine the results of the new trial with those of other trials in a meta-analysis. In its equivalence trial (1) and its systematic review (4), WHO has done both. The meta-analysis comparing oral misoprostol 600 µg with other uterotonics showed an increased relative risk (RR) of 1.42 (95% confidence interval (CI) 1.33 - 1.51) of blood loss 500 ml or more and, 1.36 (1.17 - 1.58) of blood loss 1000 ml or more. The recent update of the WHO review and another independent meta-analysis showed similar results (5).

WHO examined 31 misoprostol PPH prevention trials published between 1998 and 2001 (i.e. before the updating of the WHO systematic review) and between 2002 and 2005 (i.e. after updating the WHO systematic review). Of all of these, only four publications mention the results of the systematic reviews (2, 4) in their discussion section. Two of the four trials (1, 6) -one of which was a WHO trial- discussed the new findings in the context of the overall evidence available in the systematic review. Most of the trials do not discuss the results of all previous trials included in the systematic review and suggest equivalence between misoprostol and oxytocin in PPH prevention, which is contrary to the pooled estimate of the systematic reviews mentioned.

Unjustifiable claims of equivalence are not limited to misoprostol trials. Greene et al. (7) identified inappropriate claims of equivalence in 67% of 88 reports on the basis of a non-significant statistical comparison and only 45 of those studies were specifically aimed at studying equivalence. Journals are increasingly becoming aware of this problem. To force authors to present a summary of all existing knowledge related to a research question, the British Medical Journal (8) now requires the authors to specify in a box “what this paper adds”. Also, The Lancet recently announced that it will not consider for publication trial reports that do not refer to systematic reviews in their discussion section (9). The authors have a responsibility to discuss the results of their primary research in the context of all other existing research results in an explicit manner in the discussion sections of their research reports.

References

  • Gülmezoglu AM, Villar J, Ngoc NN, et al. for the WHO Collaborative Group To Evaluate Misoprostol in the Management of the Third Stage of Labour. WHO multicentre double-blind randomized controlled trial to evaluate the use of misoprostol in the management of the third stage of labour. The Lancet 2001;(358):689-695.
  • Gülmezoglu AM. Prostaglandins for prevention of postpartum haemorrhage. The Cochrane Database of Systematic Reviews 1997;(4):John Wiley & Sons, Ltd.
  • Gülmezoglu AM, Forna F, Villar J, et al. Prostaglandins for prevention of postpartum haemorrhage. The Cochrane Database of Systematic Reviews 2002;(3):John Wiley & Sons, Ltd.
  • Villar J, Gulmezoglu AM, Hofmeyr GJ, et al. Systematic review of randomized controlled trials of misoprostol to prevent postpartum hemorrhage. Obstetrics and Gynecology 2002;(100):1301-1312.
  • Langenbach C. Misoprostol in preventing postpartum hemorrhage. A meta-analysis. International Journal of Gynecology and Obstetrics 2006;(92):10-18.
  • Walraven G, Blum J, Dampha Y et al. Misoprostol in the management of the third stage of labour in the home delivery setting in rural Gambia: a randomised controlled trial. British Journal of Obstetrics and Gynaecology 2005;(112):1277-1283.
  • Greene WL, Concato J and Feinstein AR. Claims of equivalence in medical research: Are they supported by the evidence. Annals of Internal Medicine 2000;(132):715-722.
  • BMJ paper styles available at: http://bmj.bmjjournals.com/advice/sections.shtml. . British Medical Journal. ;():Accessed on 10 March 2006.
  • Young C and Horton R. Putting clinical trials into context. The Lancet 2005;(366):107-108.

This document should be cited as: Gülmezoglu AM, Villar J, Khanna J, Schulz KF, Lumbiganon P, Mittal S, Hofmeyr GJ, Cheng L. New research should be presented in the context of existing literature: the case of misoprostol third stage of labour trials: The WHO Reproductive Health Library, No 9; Geneva: World Health Organization.