Oxytocin receptor antagonists for inhibiting preterm labour: a Cochrane systematic review

Responses to key questions

My understanding is that TRACTOCILE is more effective than placebo for delaying preterm labour. Doesn’t this review contradict this?

The systematic review concludes that ‘…atosiban was shown to have similar tocolytic efficacy to placebo.’ However, there are a number of problems with this statement:

  • This conclusion is based on the results of a single study, which had several critical design limitations [Goodwin et al 1994]. Most importantly, women in the study received a sub-therapeutic dose of TRACTOCILE (300g/min for only 2hours with no bolus loading dose). Secondly, the study was designed to evaluate the effect of TRACTOCILE on preterm uterine activity, not efficacy. Thirdly, the gestational ages varied considerably between the different centres and went down as low as 20–23 weeks; which is both outside the current licence for TRACTOCILE and outside limits of viability at the time. Lastly, ‘delivery within 48hours of treatment’ is not the most clinically relevant indicator of efficacy. As such, it is debatable whether the Goodwin et al 1994 study should have been included in this review and this view was supported by a leading independent expert in the field[1]
  • The much larger placebo-controlled trial [n=501; Romero et al 2000] has been omitted fromthe evaluation of efficacy in this review. This trial uses the much more clinically relevant composite endpoint of non-delivery within 48hours without the need for an alternative tocolytic and demonstrates a significant difference in favour of TRACTOCILE (P=0.008).

TRACTOCILE seems to be no better than -agonists so why should I use it?

This systematic review does conclude that ‘…atosiban has similar tocolytic efficacy to -agonists.’ However, again there are a number of misrepresentations in this statement:

  • TRACTOCILE is as effective as -agonists at 48hours and more effective at 7days [Moutquin et al 2001]
  • TRACTOCILE’s comparable or superior efficacy is accompanied by a significantly superior side effects profile, particularly with respect to maternal cardiovascular events [Moutquin et al 2001]
  • Whereas the three -agonist studies are appropriately included, the inclusion of the Phase II dose ranging study by Goodwin et al in this evaluation is questionable [Goodwin et al 1996] as women in three out of four of the treatment arms received sub-therapeutic doses of TRACTOCILE. For the analysis in the review, data from all four arms are pooled, which is scientifically flawed

I have always thought that TRACTOCILE is well tolerated but this review suggests its use is associated with an increase in infant mortality compared with placebo

This review does report that ‘…when compared with placebo, TRACTOCILE resulted in an increase in infant deaths at 12months of age’. Once again, however, this statement is misleading for a number of reasons:

  • This conclusion is based on the results of only one trial [Romero et al 2000], which is acknowledged to have an imbalance in the treatment groups caused at randomisation. This resulted in a higher proportion of women at extremely early gestational ages (p=0.008) and more advanced stages of PTL (higher risk women), being treated with TRACTOCILE

Gestational age at admissionversus placebo

Gestational age at admission / Number
Atosiban / Placebo
<26 weeks / 24/246 (10%) / 13/255 (5%)
26 to <28 weeks / 19/246 (8%) / 21/255 (8%)
28 to <32 weeks / 107/246 (44%) / 105/255 (41%)
32 weeks / 96/246 (39%) / 116/255 (45%)

P=0.008 overall vs placebo

  • In the Romero et al 2000 trial, the infant deaths were determined to be associated with extreme prematurity and not study drug [Romero et al 2000], with > three quarters occurring at <26 weeks gestation and in most cases with infants weighing <650g. Babies born at 26 weeks are known to have a 1-year survival rate of just 1% [insert reference].
  • The number of infant deaths in the Romero et al 2000 trial was not significantly different from placebo – significance is only achieved in the 2-year follow-up study [Goodwin et al 1998], that includesinfant deaths up to 12 months
  • The real message from theRomero et al 2000 study should be that despite this imbalance, the systematic review found no significant differences in any of the safety data compared with placebo (including perinatal death, fetal death, neonatal death)
  • It is also of note that this trial included women at very early gestational ages (between 20–23 weeks), which is both outside the licence for TRACTOCILE and outside the accepted limits of viability at the time the trials were performed
  • In more recent, larger and well-controlled trials perinatal mortality was no higher in patients receiving TRACTOCILE [Moutquin et al 2001]; a finding that has been confirmed in a study of more than 800 women receiving TRACTOCILE in every day clinical practice [Husslein et al 2005]
  • The benefit of corticosteroid administration on infant outcome was not well understood and was not routine when some of these trials were performed and could have contributed to poorer infant outcomes

From this review it appears that TRACTOCILE is associated with lower birth weights and more maternal adverse drug reactions than placebo

This review does state that ‘…TRACTOCILE resulted in lower birth weight and an increase in maternal adverse drug reactions compared with placebo’ and ‘Compared with -agonists, atosiban increased the numbers of infants born under 1500g’. But this does not tell the whole story:

  • The placebo-controlled trial of Romero et al 2000 demonstrated comparable birth weights following TRACTOCILE and placebo
  • The statement on birthweight vs -agonists is based on two trials [Moutquinet al 2000; Goodwin et al 1996]. Since all three arms of the CAP-001 trial had an identical study protocol, and birth weight data are available from each, it is unclear why data from the salbutamol [Cabrol et al 2001] and terbutaline [Maršál et al 2001] arms have been excluded in this analysis
  • The largest randomised controlled tocolytic trials performed to date [Moutquin et al 2001] have demonstrated comparable birth weights in infants born to mothers given TRACTOCILE or -agonists (P=0.58)
  • The results of these highly controlled studies have been confirmed in a clinical use trial in more than 800 women, where significantly fewer maternal (P0.01) and fetal (P0.03) treatment-emergent adverse events were reported with TRACTOCILE compared with usual care[2] [Husslein et al 2005].

From this review it doesn’t appear that TRACTOCILE is any better than -agonists or placebo in terms of tocolytic efficacy or infant outcomes

  • In the largest tocolytic trials performed to date, the composite endpoint used demonstrated the comparable efficacy and superior safety of TRACTOCILE compared with -agonists at 48hours. At 7days both efficacy and safety were superior to -agonists
  • Recent RCOG guidelines have recommended the use of TRACTOCILE or nifedipine based on their comparable efficacy and superior safety compared with -agonists [RCOG, 2002]
  • In a large, placebo-controlled study, significantly more women receiving TRACTOCILE remained undelivered and did not receive an alternative tocolytic after 48hours (P=0.008) and 7days (P =0.003) [Romero et al 2000]
  • No tocolytic trials to date have been designed to show improved neonatal outcome

Without a direct comparison of TRACTOCILE and calcium channel blockers, what evidence is available to support the use of one drug over the other?

The review authors stress the need for a direct comparison of TRACTOCILE and calcium channel blockers. However, two recent studies have compared the most commonly used calcium channel blocker, nifedipine, and TRACTOCILE for the treatment of preterm labour. The results of these studies are in agreement with each other and demonstrate that:

  • Whereas TRACTOCILE and nifedipine demonstrated similar efficacies, TRACTOCILE was associated with significantly fewer adverse events [Kashanian et al 2005; Al-Omari et al 2004]
  • In particular,in one study [Kashanian et al 2005]most of the adverse events (17/29; 58.6%) reported in women receiving nifedipine were cardiovascular (hypotension [n=11]; palpitation [n=3]; tachycardia [n=3]) compared with none in women receiving TRACTOCILE

Recent systematic reviews and meta-analyses have suggested the use of calcium channel blockers instead of other tocolytics based on superior efficacy and safety

  • These recommendations have been based on meta-analyses and systematic reviews of generally small, poorly-conducted studies [Lamont et al 2005]
  • More recently, a number of studies have been published documenting serious maternal cardiovascular and pulmonary adverse events [Hodges et al 2004; Oei et al 1999; Verhaert and van Acker 2004; Bal et al 2004; Vaast et al 2004]
  • A case of fetal death has also been reported following the tocolytic administration of nifedipine [van Veen et al 2005]
  • Nifedipine remains unlicensed for use in pregnancy. Its efficacy and safety has never been demonstrated in pregnant women within the context of licensing trials. Indeed the authors of this systematic review state that there is ‘no evidence that either of them[nifedipine and atosiban]are clinically superior to placebo when it comes to important perinatal outcome including long-term safety for babies’
  • Two recent studies have directly compared TRACTOCILE and nifedipine in a clinical situation, and both demonstrated that the two drugs have comparable efficacy in delaying PTL, but that TRACTOCILE is associated with significantly fewer maternal adverse events in 71 and 80 women, respectively [Al-Omari et al 2004; Kashanian et al 2005].
  • Maternal adverse events in the Kashanian et al 2005 trial occurred in fewer women receiving TRACTOCILE (17.5%) compared with nifedipine (40%; P=0.027). Most of the adverse events (17/29; 58.6%) reported in women receiving nifedipine were cardiovascular (hypotension [n=11]; palpitation [n=3]; tachycardia [n=3]) compared with none in women receiving TRACTOCILE
  • There has never been a trial of nifedipine vs placebo

Further well-designed randomised controlled trials should address the relative risks and benefits of TRACTOCILE, nifedipine and placebo

  • The manufacturers of nifedipine have actively discouraged the use of nifedipine in preterm labour and maintain that there never has been, and never will be, a registration trial of nifedipine in tocolysis [Lamont, 2004]
  • Of the [new] tocolytics registered, TRACTOCILE has undergone the most thorough pharmacological and methodological evaluation [Tsatsaris et al 2004]
  • Two recent studies have directly compared TRACTOCILE and nifedipine in a clinical situation, and both demonstrated that the two drugs have comparable efficacy in delaying PTL, but that TRACTOCILE is associated with significantly fewer maternal adverse events in 71 and 80 women, respectively [Al-Omari et al 2004; Kashanian et al 2005].

This review points to a number of issues raised by the FDA surrounding the use of TRACTOCILE. Should I be concerned?

During the early stages of the development of TRACTOCILE in the mid-1990s the FDA did suggest that the gestation-dependent expression of myometrial oxytocin receptors may make TRACTOCILE more appropriate for use at later gestational ages. However:

  • There has never been any in-vitro or in-vivo evidence to support this concept and the trial investigators dismissed this proposition at the time
  • A recent trial has demonstrated that TRACTOCILE is effective in the management of preterm labour at low gestational ages (18–24weeks) and exhibited its established favourable side effects profile [Richter et al 2005]

The review also reports increased infant mortality and lower birth weight with TRACTOCILE compared with placebo. The FDA has suggested that these adverse infant outcomes could be due to fetal vasopressin receptor blockade by TRACTOCILE. However:

  • This theory was largely discredited at the time as there was no evidence at all to support it
  • A subsequent animal study investigated the potential anti-vasopressinergic effects of TRACTOCILE on fetal sheep. The study demonstrated that TRACTOCILE had no long-term effects on fetal oxygenation [Nathanieisz P, data on file]
  • Animal studies have also demonstrated that fetal plasma levels of atosiban are at or below the detection limit and that fetal blood pressure is not affected [Greig et al 1993]
  • A much more plausible explanation for the increased infant mortality and lower birth weight with TRACTOCILE is the imbalance in treatment groups in the study by Romero et al 2000, which is acknowledged by the wider obstetric community
  • In fact it is noteworthy that despite this imbalance, most of the mortality data in this trial (perinatal death, fetal death, neonatal death) were comparable with placebo. Even the infant mortality rate is only significantly different in the follow-up study [Goodwin et al 1998], with the inclusion of infant deaths up to 12months

Cochrane is a respected organisation well known for the publication of evidence-based systematic reviews. The findings of this review must be accurate.

Cochrane has justifiably earned its reputation in evidence-based medicine by the publication of high quality science. However, as with all large organisations it is not infallible.

In order to maintain the impartiality and independence for which the Cochrane Library is respected, it is critical that there is balanced opinion among the authors of every review. However, in this case, the independence of some of the authors might be debatable. It is worth noting some key points:

  • The authors of this Cochrane review are self-appointed
  • Drs Papatsonis and Flenady have both published prolifically on calcium channel blockers (including the recent Cochrane Systematic Review on calcium channel blockers for use in preterm labour) and have very strongly advocated the use of nifedipine in this indication
  • As there is no mention of TRACTOCILE in the previous Cochrane Systematic Review on calcium channel blockers for use in preterm labour, it is unclear why nifedipine data are specifically mentioned in this review. Again, it might be argued by some readers that this is reflective of the authors’ own clinical preferences

References

  1. Papatsonis D, Flenady V, Cole S and Liley H. Oxytocin receptor antagonists for inhibiting preterm labour: a Cochrane systematic review. Cochrane Database Syst Rev 2005; Jul 20(3): CD004452.
  2. Al-Omari WR. Atosiban and nifedipine in acute tocolysis, comparative study [abstract]. XVIII European Congress of Obstetrics and Gynaecology; 2004 May 12–15; Athens, Greece. 2004: 103.
  3. Bal L, Thierry S, Brocas E, Adam M, Van de Louw A, Tenaillon A. Pulmonary edema induced by calcium-channel blockade for tocolysis. Anesth Analg 2004; 99(3): 910–911.
  4. Cabrol D and the French/Australian Atosiban Investigators Group. Treatment of preterm labor with the oxytocin antagonist atosiban: a double-blind, randomized, controlled comparison with salbutamol. Eur J Obstet Gynecol Reprod Biol 2001; 98(2): 177–185.
  5. Goodwin TM, Paul R, Silver H, Spellacy W, Parsons M, Chez R, Hayashi R, Valenzuela G, Creasy GW, Merriman R. The effect of the oxytocin antagonist atosiban on preterm uterine activity in the human. Am J Obstet Gynecol 1994; 170(2): 474–478.
  6. Goodwin TM, Valenzuela GJ, Silver H, Creasy G. Dose ranging study of the oxytocin antagonist atosiban in the treatment of preterm labor. Atosiban Study Group. Obstet Gynecol 1996; 88(3): 331–336.
  7. Goodwin TM, Randall H, Perry K, Menard MK, Bauer C, Shangold G et al. A report on infant outcomes at 6 and 12months after the use of atosiban in the management of preterm labour. 46th ACOG Annual Meeting; 1998 May 9–13; New Orleans, Louisiana.
  8. Greig PC, MassmannGA, Demarest KT, Weglein RC, Holland ML, Figueroa JP. Maternal and fetal cardiovascular effects and placental transfer of the oxytocin antagonist atosiban in late-gestation pregnant sheep.Am J Obstet Gynecol 1993;169(4):897–902.
  9. Hodges R, Barkehall-Thomas A, Tippett C. Maternal hypoxia associated with nifedipine for threatened preterm labour. BJOG 2004; 111(4): 380–381.
  10. Husslein et al. Poster presented at COGI, Athens, Greece, 2005.
  11. Kashanian M, Akbarian AR, Soltanzadeh M. Atosiban and nifedipin for the treatment of preterm labor.Int J Gynaecol Obstet 2005; [epub before print].
  12. Lamont RF. Reply to article by King et al. Aust N Z J Obstet Gynaecol 2004; 44(3): 275–276.
  13. Maršál K and the European Atosiban Study Group. The oxytocin antagonist atosiban versus the beta-agonist terbutaline in the treatment of preterm labor. A randomized, double-blind, controlled study. Acta Obstet Gynecol Scand 2001; 80(5): 413–422.
  14. Moutquin J-M, Sherman D, Cohen H, Mohide PT, Hochner-Celnikier D, Fejgin M, Liston RM, Dansereau J, Mazor M, Shalev E, Boucher M, Glezerman M, Zimmer EZ, Rabinovici J. Double-blind, randomized, controlled trial of atosiban and ritodrine in the treatment of preterm labor: a multicenter effectiveness and safety study. Am J Obstet Gynecol 2000; 182(5): 1191–1199.
  15. Moutquin J-M and the Worldwide Atosiban versus Beta-agonists Study Group. Effectiveness and safety of the oxytocin antagonist atosiban versus beta-adrenergic agonists in the treatment of preterm labour. The Worldwide Atosiban versus Beta-agonists Study Group. BJOG 2001; 108(2): 133–142.
  16. Oei SG, OeiSK, Brolmann HA. Myocardial infarction during nifedipine therapy for preterm labor. N Engl J Med 1999; 340(2):154.
  17. RichterON, Dorn C, van de Vondel P, Ulrich U, Schmolling J. Tocolysis with atosiban: experience in the management of premature labor before 24 weeks of pregnancy.Arch Gynecol Obstet2005;272(1):26–30.
  18. Romero R, Sibai BM, Sanchez-Ramos L, Valenzuela GJ, Veille JC, Tabor B, Perry KG, Varner M, Goodwin TM, Lane R, Smith J, Shangold G, Creasy GW. An oxytocin receptor antagonist (atosiban) in the treatment of preterm labor: a randomized, double-blind, placebo-controlled trial with tocolytic rescue. Am J Obstet Gynecol 2000; 182(5): 1173–1183.
  19. RCOG guidelines, 2002.
  20. Tsatsaris V, Carbonne B, Cabrol D. Atosiban for preterm labour. Drugs 2004; 64(4): 375–382.
  21. Vaast P, Dubreucq-Fossaert S, Houfflin-Debarge V, Provost-Helou N, Ducloy-Bouthors AS, Puech F, Subtil D. Acute pulmonary oedema during nicardipine therapy for premature labour; Report of five cases. Eur J Obstet Gynecol Reprod Biol 2004; 113(1): 98–99.
  22. Valenzuela GJ, Sanchez-Ramos L, Romero R, Silver HM, Koltun WD, Millar L, Hobbins J, Rayburn W, Shangold G, Wang J, Smith J, Creasy GW. Maintenance treatment of preterm labor with the oxytocin antagonist atosiban. The Atosiban PTL-098 Study Group. Am J Obstet Gynecol 2000; 182(5): 1184–1190.
  23. Van Veen AJ, Pelinck MJ, Pampus MG, Erwich JJ. Severe hypotension and fetal death due to tocolysis with nifedipine. BJOG 2005; 112(4): 509–510.
  24. Verhaert D, Van Acker R. Acute myocardial infarction during pregnancy. Acta Cardiol 2004; 59(3): 331–339.

Appendix

Summary of trials included in the Cochrane review of oxytocin antagonists

Trial / Description / Patients
Atosiban versus placebo
Goodwin et al 1994 /
  • Phase II study to assess effect of atosibanon preterm uterine activity (not efficacy)
  • Atosibanadministered at sub-therapeutic dose – 300g/min for only 2hours, with no bolus loading dose
  • Women unlikely to be in true preterm labour
/ 120
Romero et al 2000 /
  • Phase III trial (PTL-096)
  • Imbalance of groups at randomisation; with more women at lower gestational ages (P=0.008) and more advanced labour in the atosiban group
/ 501
Atosiban versus -agonists
Goodwin et al 1996 /
  • Phase II dose ranging study of atosiban vs ritodrine
  • Four atosiban dosing arms pooled (6.5mg bolus + 300g/min; placebo bolus + 300g/min; 2mg bolus + 100g/min; 0.6mg bolus + 30g/min) – but at least three arms were sub-therapeutic
/ 302
Moutquin et al 2000 /
  • Phase III trial – CAP-001; atosibanvs ritodrine
/ 247
Cabrol et al 2001 /
  • Phase III trial – CAP-001; atosibanvs salbutamol
/ 241
Maršál et al 2001 /
  • Phase III trial – CAP-001; atosibanvs terbutaline
/ 249

1

F/002/09/05

[1]Douglas Badenoch, Minervation Ltd, Oxford, UK.

[2]Usual care included -agonists, calcium channel blockers and bed rest