Adverse effects of tocolytic therapy

Transcription

1 BJOG: an International Journal of Obstetrics and Gynaecology March 2005, Vol. 112, Supplement 1, pp Adverse effects of tocolytic therapy Steve Caritis The rationale for using tocolytics in preterm labour is to enable transfer of the mother to a tertiary centre and to prolong pregnancy sufficiently so that glucocorticoids can be administered to the mother. There is little question that these short term objectives can be achieved with contemporary tocolytics. Whether tocolytics can maintain pregnancy for sufficient periods to enable in utero maturation to occur remains an unresolved question. When a decision is made to use tocolytics, the clinician is faced with a multitude of choices with side effects, efficacy and ease of administration generally being the most important considerations. Placebocontrolled studies suggest that the h-agonists, prostaglandin inhibitors and atosiban are effective in prolonging pregnancy for hours. Of these three agents, atosiban has the best safety profile. There are no placebocontrolled studies with calcium channel blockers or nitric oxide donors. However, because of their ease of use and efficacy compared with the h-agonists, calcium channel blockers are widely used. Calcium channel blockers appear to have a better safety profile than the h-agonists, but there are still significant cardiovascular side effects associated with their use. Indomethacin, although proven to be efficacious, has a safety profile that limits its utility for other than short courses. Magnesium sulphate is the most commonly used tocolytic in the United States, despite a lack of evidence for its efficacy. Although magnesium sulphate appears to have a good safety profile, serious side effects have been reported with its use. The choice of tocolytics is commonly based on personal preference. Whichever tocolytic is chosen, the fundamental parturitional process is not reversed by contemporary treatment, rather a reduction in uterine response to a stimulant; thus, the expectations of tocolytic treatment need to be reconsidered. INTRODUCTION Spontaneous preterm labour and birth, with or without rupture of the fetal membranes, accounts for 9% of all births. The number of women presenting with signs and symptoms of preterm labour is roughly double that amount. Of approximately four million pregnancies annually in the United States, 720,000 women experience an episode of presumed preterm labour and may receive tocolytics. The purpose of treatment may be to allow for transport to a tertiary centre, to allow for administration of glucocorticoids to enhance fetal pulmonary maturity or to attempt to prolong the pregnancy sufficiently to allow for additional fetal maturity. 1,2 A multitude of tocolytics acting through diverse pathways has been used to treat preterm labour, but only two agents, atosiban and ritodrine, were developed specifically for preterm labour inhibition. The remaining agents have been found to inhibit myometrial contractility to some extent even though this may not have been the primary reason for their development or marketing. These agents have been used in combination or sequentially for the inhibition of labour. Unfortunately, life-threatening side effects have Department of Obstetrics, Gynecology, and Reproductive Sciences, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania, USA Correspondence: Professor S. Caritis, 300 Halket Street, Pittsburgh, Pennsylvania 15213, USA. D RCOG 2005 BJOG: an International Journal of Obstetrics and Gynaecology been reported with some of these medications. 3,4 In this review, we will address the issue of adverse effects of tocolytics and examine some factors that contribute to these adverse effects. We will also discuss the most commonly used tocolytics, including h-adrenergic receptor agonists, cyclooxygenase (COX) inhibitors, calcium channel blockers, nitric oxide donors, oxytocin receptor antagonists and magnesium sulphate. For the purpose of this review, all effects other than the physiological effects on the myometrium will be considered as adverse effects. This includes physiological effects in non-myometrial tissues, non-physiological effects on myometrial or non-myometrial tissues and pathological effects. FACTORS DETERMINING LIKELIHOOD OF ADVERSE EFFECTS Some factors relevant in determining the likelihood of adverse effects with commonly used tocolytics are listed in Table 1. The tocolytics currently in use affect myometrial contractility by one of two major pathways (Fig. 1). They affect either the contractile proteins (usually the phosphorylation of myosin) by generation or alteration of an intracellular messenger or they inhibit the synthesis of, or block the action of, a known myometrial stimulant. 1 Among the first group are the h-adrenergic receptor agonists, nitric oxide donors, magnesium sulphate and the calcium channel blockers. The h-adrenergic receptor agents, by binding with

2 ADVERSE EFFECTS OF TOCOLYTIC THERAPY 75 Table 1. Factors determining likelihood of side effects. 1. Mechanism of drug action 2. Target tissues 3. Dosage and route of administration 4. Pharmacokinetics and pharmacodynamics cell membrane h-adrenergic receptors, lead to an increase in intracellular cyclic AMP that inhibits the ability of myosin light chain kinase to phosphorylate myosin. 5,6 Nitric oxide donors produce a similar response through cyclic GMP. 1 Magnesium sulphate and calcium channel blockers lower intracellular calcium, thereby reducing the activity of myosin light chain kinase and inhibiting the phosphorylation of myosin. 1,6,7 Oxytocin and prostaglandins are the major endogenous myometrial stimulants. Atosiban is an oxytocin receptor antagonist. Binding to the receptor in the myometrium and other gestational tissues prevents the generation of inositol triphosphate, the messenger that increases intracellular calcium and causes myometrial contractions and up-regulates prostaglandin production. 1,8,9 Prostaglandins are produced in the myometrium and other gestational tissues. Agents such as indomethacin inhibit the COX enzymes, which are key to the production of these prostaglandins. General inhibitors of both the COX-1 (constitutive) and COX-2 (inducible) enzymes include indomethacin and sulindac. Selective COX-2 inhibitors include agents such as nimesulide and rofecoxib. 1,10 PHYSIOLOGICAL EFFECTS OF LABOUR-INHIBITING AGENTS When tocolytics are administered to a woman in preterm labour, the primary target is the myometrial cell, although the cellular effects of these agents are not limited to the myometrial cell. The distribution of tissues responsive to the cellular effects of these agents is another major determinant of the likelihood and type of side effect seen (Table 2). Table 2 shows some of the tissues responsive to the most commonly used labour-inhibiting agents. The list is not intended to be exhaustive, but to provide some insight into the non-myometrial effects which might be expected with each agent. The wide distribution of responsive tissue other than the uterus dictates that numerous undesired side effects are possible with most of these agents. 1,11 16 The h-adrenergic receptor agonists have the most numerous tissue effects with localisation in both the cardiovascular and metabolic systems. This results in both cardiovascular and metabolic side effects, regardless of the specificity of the agonist for the h-1 or the h-2 receptor. The ability to generate intracellular signals may differ between tissues and this relationship is critical in determining the probability of side effects. For the h-adrenergic receptor agonists, clinical experience has shown that even the lowest intravenous dose needed to inhibit uterine contractions will elicit an increase in heart rate resulting in cardiovascular symptomatology whenever these agents are used. Agents such as nitric oxide donors and calcium channel blockers also have prominent effects on vascular smooth muscle, and clinical experience likewise has shown that use of these agents will be associated with some cardiovascular symptomatology. Magnesium sulphate exhibits substantial effects at the neuromuscular junction resulting in symptoms of weakness and fatigue. The wide distribution of tissues able to generate prostaglandins results in a whole host of side effects with the use of COX inhibitors. Although it would seem that specific COX-2 enzyme inhibitors should have fewer side effects, this has not proven to be the case, especially for fetal ductal closure and reduced renal blood flow. 1,10 Most striking in Table 2 is the limited tissues expressing oxytocin receptors. Fig. 1. Site of action of tocolytic agents. D RCOG 2005 BJOG: an International Journal of Obstetrics and Gynaecology 112 (Suppl. 1), pp

3 76 S. CARITIS Table 2. Non-myometrial tissues responsive to several tocolytics. Agent Tissues Effect h-adrenergic receptor agonists Myocardium h-1 increased chrono/isotropism Smooth muscles h-2 vasodilation, bronchodilation, decreased gut motility, ureteral dilation Juxtaglomerular cells h-1 increased renin Skeletal muscle h-2 glycogenolysis, increased K þ uptake Liver h-2 gluconeogenesis, glycogenolysis COX inhibitors Vascular smooth muscle Vasoconstriction/vasodilation Stomach Alteration in gastric acid release and mucosal barrier Platelets Inhibition of thromboxane Kidney Regulation of renal blood flow, GFR and renin Oxytocin antagonists Breast myoepithelial cell Inhibition of milk letdown Calcium channel blockers Vascular smooth muscle Vasodilation Cardiac conduction Inhibition of electrical conductivity Magnesium sulphate Smooth muscle Relaxation Brain NMDA receptor Sleepiness Neuromuscular junction blockage Muscle weakness Nitric oxide donors Vascular smooth muscle Relaxation This would seem to represent an optimal circumstance in terms of side effects and has proven to be the case with atosiban. 1,8,9 NON-PHYSIOLOGICAL/ PATHOLOGICAL EFFECTS OF LABOUR-INHIBITING AGENTS The agents used to inhibit preterm labour may elicit nonreceptor-mediated responses in tissues, leading to undesirable side effects. In the same way, receptor-mediated effects can be elicited, which are excessive and lead to tissue injury or other pathological effects. Pathological effects might also occur as a result of non-physiological effects in various tissues. Examples of non-specific effects include the nausea associated with atosiban or perhaps magnesium sulphate, and the oedema noted with calcium channel blocker usage. Examples of pathological effects include the oligohydramnios and fetal ductal closure associated with indomethacin and the myocardial ischaemia or pulmonary oedema noted with several different agents. The adverse effects of pulmonary oedema and myocardial ischaemia merit additional commentary. Pulmonary oedema associated with tocolytic therapy has been reported with h-agonists, magnesium sulphate and the calcium channel blockers. 14 This complication generally results as a consequence of several cofactors. Most important is volume overload, which is commonly seen with excess fluid administration, especially fluid with salt. The h-agonists are particularly problematic because of the h-adrenergic receptor-mediated effects on renin secretion and arginine vasopressin release, which increase sodium and water retention. Pulmonary oedema has also been associated with blood transfusion and multifetal gestation, and other conditions in which blood volume is expanded. 2 Additional factors beside volume overload are generally required to induce pulmonary oedema in otherwise healthy women. Injury to the endothelial cells of the pulmonary vasculature either by infection (intra-amniotic, renal, pulmonary) or pre-eclampsia is probably the most common and potent aetiological cofactor in cases of pulmonary oedema. This pathologic condition is not unique to h-agonists but can occur with any labour-inhibiting drug in the circumstances described above. 2,3 Myocardial ischaemia and tissue injury have been noted with h-agonists, magnesium and with calcium channel blockers. 2,3 These effects are probably attributable to a combination of increased myocardial oxygen consumption and decreased myocardial oxygen delivery. The former occurs as a consequence of the increased heart rate, stroke volume and cardiac output while the latter may be due to reduced coronary blood flow secondary to systemic hypotension. Theoretically, any agent capable of causing the cardiovascular changes defined above could lead to myocardial ischaemia. 3 Fig. 2. Relationship of ritodrine concentration and side effects. Infusion rate in closed circles in Ag/min, ritodrine concentration in ng/ml in open circles. SOB ¼ shortness of breath; CP ¼ chest pain. D RCOG 2005 BJOG: an International Journal of Obstetrics and Gynaecology 112 (Suppl. 1), pp

4 ADVERSE EFFECTS OF TOCOLYTIC THERAPY 77 Table 3. A comparison of ritodrine administered intravenously or intramuscularly. Side effects Intravenous (N ¼ 49) Intramuscular (N ¼ 49) Chest pain 9 2 Shortness of breath 4 5 Vomiting 8 3 DBP 40 mmhg MHR >130 mmhg 18 6 Patients with 1 of above 34 (69%)* 21 (43%) Failed therapy 16 (33%) 16 (33%) DBP ¼ diastolic blood pressure; MHR ¼ maternal heart rate. * <0.05. DOSE AND ROUTE OF ADMINISTRATION AS A DETERMINANT OF SIDE EFFECTS There is a clear relationship between dose and plasma concentration and the risk of side effects. This is most evident with agents exhibiting cardiovascular side effects. Figure 2 depicts data from a patient receiving ritodrine. In two instances, when plasma concentration rose acutely, a side effect occurred. These effects appear to be more prominent when the medications are administered intravenously compared with intramuscular administration. Table 3 compares the risk of side effects in women randomised to receive ritodrine intravenously or intramuscularly (unpublished data). In this study, side effects were significantly less common after intramuscular administration despite comparable efficacy. adequate dose is used. Conversely, we have shown that inadequate pharmacokinetic data have led to excessive plasma concentrations of intravenously administered ritodrine because the recommended infusion protocol leads to an excess in the necessary plasma concentrations required for inhibition. 18 Pharmacoepidemiological studies have not been performed for most commonly used labour-inhibiting agents. The genetic variation in drug-metabolising enzymes plays a major role in individual responses to any pharmacological agent. As depicted in Fig. 3, there is a very large variation in steady state concentrations in women receiving identical doses of ritodrine intravenously. This variation is not reduced by corrections for gestational age or body mass index (BMI) but is attributable to a large variation in hepatic conjugation. 18 Pharmacodynamic studies relate end organ response to plasma concentrations. Such studies are generally performed early in the process of drug development. Because most labour-inhibiting drugs were not approved for preterm labour, adequate pharmacodynamic data are commonly lacking. Such studies usually establish a dose range, both an upper and lower limit. A measurable end organ response (i.e. uterine contractions) is needed for such studies. The uterine response is commonly recorded non-invasively and is therefore not well quantified. Furthermore, the uterine response is commonly measured in women who are not in true preterm labour but rather may be experiencing false labour. The dose of labour-inhibiting agent required to stop contractions not associated with an activated myometrium (increased gap junctions, increased oxytocin receptors and enhanced contractility) is likely to be lower than the dose required to inhibit a contracting uterus that has undergone PHARMACOKINETICS AND PHARMACODYNAMICS Only ritodrine and atosiban were developed specifically for labour inhibition, whereas the rest of the agents currently in use have been developed or used primarily for other indications. Consequently, inadequate pharmacological information exists for most of these agents when they are used in pregnancy to inhibit preterm labour. Even when approval for use in preterm labour is granted by a federal agency, adequate pharmacokinetic and pharmacodynamic information may not be obtained. For example, when ritodrine was first approved, all pharmacological data regarding its kinetic behaviour were obtained in men and non-pregnant women. This approach is now known to be problematic because the metabolism, distribution, absorption and elimination of most drugs are altered during pregnancy. Studies in animals and humans have shown that an oral dose of ritodrine or terbutaline achieves far lower concentrations in pregnant than in non-pregnant subjects. In this way, the maximum oral dose of ritodrine recommended (i.e. 120 mg/day) is far too low to achieve therapeutic concentrations. 17 Consequently, the agent has seemed to be ineffective when in fact it may be effective if an Fig. 3. Variation in plasma concentration with an infusion of 50 Ag/min ritodrine. Simulated plasma concentrations in 13 women receiving a constant infusion of ritodrine 50 Ag/min. Concentrations based on each individual pharmacokinetic parameters. Reprinted from Am J Obstet Gynecol, 162, Caritis S, Venkataramanan R, Darby M, Chiao JP, Krew M. Pharmacokinetics of ritodrine administered intravenously: recommendations for changes in the current regimen, pp Copyright 1990, with permission from Elsevier. D RCOG 2005 BJOG: an International Journal of Obstetrics and Gynaecology 112 (Suppl. 1), pp

5 78 S. CARITIS activation. In such circumstances, the lower dose limit may be too low. Alternatively, when establishing an upper dosing range, no clear endpoint exists. The upper dose range of terbutaline administered intravenously ranges from 17.5 mg/ml at our institution to 80 Ag/mL at other institutions. 19 The lack of pharmacokinetic and pharmacodynamic data has led to many problems with labour-inhibiting drugs, many of which are avoidable. The off-label use of these agents and the unwillingness of the pharmaceutical industry to consider preterm labour treatment as a therapeutic indication for existing products are at the root of the problem with these agents. FUTURE DIRECTIONS The ideal labour-inhibiting agents would exclusively target the myometrium or the labour stimulus itself. It would be orally active with a rapid onset of action and long duration. The agent would be devoid of maternal and fetal side effects and would be highly effective. Such an agent does not currently exist. Until the factor(s) that initiate the parturitional process is (are) identified, it is not likely that a highly effective agent will be developed. The alternative strategy has been to block the end organ response (i.e. to inhibit myometrial contractility without regard to the initiating factors). Because oxytocin and prostaglandins appear to be the myometrial stimulants of the final common pathway of labour, blockade or suppression of these stimulants may prove to be sufficient to achieve at least some of the goals of tocolytic therapy. If the initiating factor for preterm labour is infection, then the best course of action might be delivery, not contraction suppression. References 1. Simhan H, Caritis S. Inhibition of preterm labor. Up-to-Date CD-ROM 2004 (September). 2. Jeyabalan A, Caritis S. Pharmacologic inhibition of preterm labor. Clin Obstet Gynecol 2002;45: Hankins GDV. Complications of beta-sympathomimetic tocolytic agents. In: Clark SL, Cotton DB, Hankins GDV, Phelan JP. Critical Care Obstetrics, 2nd edition. Boston: Blackwell Scientific Publications, 1991: Vaast P, Dubreucq-Fossaert S, Houfflin-Debarge V, et al. Acute pulmonary oedema during nicardipine therapy for premature labour; report of five cases. Eur J Obstet Gynecol Reprod Biol 2004;113: Boyle J. Beta-adrenergic agonists. Clinic Obstet Gynecol 1995;38: Idama TO, Lindow SW. Magnesium sulphate: a review of clinical pharmacology applied to obstetrics. Br J Obstet Gynaecol 1998;105: Ray D, Dyson D. Calcium channel blockers. Clin Obstet Gynecol 1995;38: The Worldwide Atosiban vs. Beta-agonists Study Group. Effectiveness and safety of the oxytocin antagonist atosiban versus betaadrenergic agonists in the treatment of preterm labour. Br J Obstet Gynaecol 2001;108: Romero R, Sibai BM, Sanchez-Ramos L, et al. 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: Gordon M, Samuels P. Indomethacin. Clinic Obstet Gynecol 1995; 38: Hill W. Risks and complications of tocolysis. Clinic Obstet Gynecol 1995;38: Benedetti T. Maternal complications of parenteral h-sympathomimetic therapy for premature labor. Am J Obstet Gynecol 1983;145: Gyetvai K, Hannah M, Hodnett E, Ohlsson A. Tocolytics for preterm labor: a systematic review. Obstet Gynecol 1999;94: Duckitt K, Thornton S. Nitric oxide donors for the treatment of preterm labour. Cochrane Database Syst Rev 2002:CD King JF, Flenady VJ, Papatsonis DM, Dekker GA, Carbonne B. Calcium channel blockers for inhibiting preterm labour. Cochrane Database Syst Rev 2003:CD Crowther C, Hiller J, Doyle L. Magnesium sulphate for preventing preterm birth in threatened preterm labour. Cochrane Database Syst Rev 2002:CD Caritis S, Venkataramanan R, Cotroneo M, Chiao JP. Pharmacokinetics of orally administered ritodrine. Am J Obstet Gynecol 1989; 161: Caritis S, Venkataramanan R, Darby M, Chiao JP, Krew M. Pharmacokinetics of ritodrine administered intravenously: recommendations for changes in the current regimen. Am J Obstet Gynecol 1990;162: Wallace R, Caldwell D, Ansbacher R, Otterson W. Inhibition of premature labor by terbutaline. Am J Obstet Gynecol 1978;51: D RCOG 2005 BJOG: an International Journal of Obstetrics and Gynaecology 112 (Suppl. 1), pp