Codeine in mothers and children: where are we now?

Abstract

The death of a child is always a tragedy, so when it was reported that a toddler had died at home in North America following ingestion of codeine prescribed for postoperative analgesia 1, clinicians were understandably alarmed. This child was subsequently found to be what is described as an ultra-rapid metaboliser of codeine. A study in children demonstrated hypercarbia and a depressed ventilatory response to carbon dioxide at plasma morphine concentrations above 20 ng.l−1 2, and his post-mortem plasma concentration of morphine was 32 ng.l−1. It was presumed that excessively high morphine levels, together with airway obstruction, caused fatal respiratory depression. There followed a further report of two deaths and a ‘near miss’ related to the use of postoperative codeine in children 3. In the ensuing months, the Food and Drug Administration (FDA) in the USA 4, the European Medicines Agency (EMA) 5 and the Medicines and Healthcare products Regulatory Agency (MHRA) 6 all issued warning notices related to the use of codeine in children. The MHRA recommended that “codeine should only be used to relieve acute moderate pain in children older than 12 years and only if it cannot be relieved by other painkillers such as paracetamol or ibuprofen”. In addition, the MHRA stated that “codeine is contraindicated in all children (i.e. younger than 18 years) who undergo tonsillectomy or adenoidectomy (or both) for obstructive sleep apnoea”, and that it “is not recommended for children whose breathing might be compromised, including those with neuromuscular disorders, severe cardiac or respiratory conditions, upper respiratory or lung infections, multiple trauma or extensive surgical procedures” 6. The MHRA provided no advice regarding children without obstructive sleep apnoea having intermediate-sized surgery such as tonsillectomy/adenoidectomy. In November 2013, a joint statement by the Royal College of Anaesthetists (RCoA), the Association of Paediatric Anaesthetists (APA) and the Royal College of Paediatrics and Child Health (RCPCH) was released 7. Their conclusions were non-committal; “Within the UK different solutions are being employed. These include continuing to use codeine with increased caution or adopting alternative opioid medication regimens: oral morphine, dihydrocodeine, oxycodone or tramadol are potential alternatives.” So what is the problem with codeine; after all, it has been in use for many years and prescribed, without apparent problems, for millions of patients worldwide? One issue is that codeine's precise mechanism of action is unclear. Seventy to eighty percent of codeine is metabolised by glucuronyltransferase enzymes in the liver to codeine-6-glucuronide, the analgesic activity of which is unknown in humans, and 10% to norcodeine, a metabolite with no analgesic properties. A small proportion (5-10%) is metabolised by the cytochrome P450 2D6 enzyme system to morphine. Because codeine has only 0.5% of the affinity for mu-opioid receptors compared with morphine, it is widely believed that the major analgesic effect of codeine is as a result of its conversion to morphine 8, 9. Codeine therefore behaves like a pro-drug, requiring metabolism to become clinically active. The bigger problem, however, is that there are a substantial number of different alleles coding for activity of the cytochrome 2D6 enzyme system and, indeed, gene duplication in a number of important ethnic groups. This results in a wide spectrum of enzyme activity (phenotype) across the range of different genotypes (genetic polymorphism). At one extreme, two abnormal genes lead to individuals with no 2D6 enzyme activity and no ability to convert codeine to morphine, so-called poor metabolisers. At the other extreme, duplication of active alleles results in individuals with unusually high 2D6 enzyme activity and extensive conversion of codeine to morphine (ultra-rapid metabolisers). A typical North European population in the UK consists of 5-10% poor metabolisers and 0.03% ultra-rapid metabolisers 10. Ultra-rapid metabolisers are particularly prevalent in Africa/Ethiopia (29%) and Saudi Arabia (21%) and this is thought to be an evolutionary development in relation to diet 11. In addition to the wide genetic variability in enzyme activity, 2D6 is also inhibited or induced by the use of concomitant medications, which may also alter codeine's efficacy or side-effect profile. Codeine remains an immensely popular analgesic during the peripartum period, with two recent studies indicating that the incidence of exposure to codeine during pregnancy is 5-6%, increasing to almost 30% in the postpartum period 12-14. To enable a comprehensive evaluation of codeine usage in pregnancy and the postpartum period, it is essential to determine its safety during the first trimester, its efficacy, and its impact on breastfeeding and the neonate. Evidence for the teratogenic effects of codeine in animal studies is inconclusive and seems to depend on the animal species tested 15, 16. Of more concern is the association between use of codeine and birth defects in human epidemiological studies. In a recent population based case-control study 17, peri-conceptual treatment with opioid analgesics, in particular codeine, was associated with a higher incidence of birth defects, in particular a twofold increased risk of cardiac defects (atrioventricular septal defect, hypoplastic left heart syndrome, aortic stenosis) and spina bifida. Other studies have reported associations between peri-conceptual use of codeine and the incidence of neural tube defects 18 and neuroblastoma 19. Co-administration of codeine with paracetamol is unlikely to explain the higher rate of congenital abnormalities, because paracetamol per se is not associated with an increase in birth abnormalities 20, 21. The worrying associations between use of codeine and birth defects are, however, tempered by a recent Norwegian cohort study that failed to identify an association between use of codeine in pregnancy and an increased rate of congenital malformations 22. Despite inconsistent evidence, there are sufficient red flags to question the safety of codeine during pregnancy. Pregnancy outcomes such as mode of delivery and immediate complications are largely determined by the wellbeing of the maternal-fetal unit, and it is extremely unlikely that they would be affected by antenatal opioid therapy. However, Nezvalova-Henriksen et al. found a significant association between the use of codeine in pregnancy and adverse obstetric outcomes 22. Specifically, the use of codeine at any time during pregnancy was associated with an increased risk of elective and emergency caesarean section and postpartum haemorrhage. Although the associations remain strong, patients who received codeine were also overweight, had multiple co-morbidities and were prescribed additional psychotropic agents, including antidepressants. Therefore, the causal role of codeine needs to be further evaluated, especially as recent research suggests that use of antidepressants during pregnancy per se is associated with an increased incidence of postpartum haemorrhage 23. Because of genetic polymorphism, there is great variability in the relative contribution of morphine to analgesia. A pilot study tested the association between the 2D6 genotype and codeine analgesia in 45 women following elective caesarean section 24. Although the study was underpowered to detect differences in analgesia between the four different genotypes, women at the genotypic extremes reported codeine effects consistent with their genotype: two poor metabolisers reported no analgesia as a result of taking codeine whereas two of the three utra-rapid metabolisers reported immediate pain relief from codeine but stopped taking it due to dizziness and constipation. Unlike specific studies targeting the codeine-genotype interaction for adverse effects, studies comparing the analgesic efficacy of codeine are limited because of its almost universal co-administration with paracetamol. Despite this apparent confounder, the majority of studies have shown no differences in analgesia between paracetamol-codeine and non-steroidal anti-inflammatory drugs (NSAIDs). In fact, a recent meta-analysis concluded that NSAIDS are as effective as paracetamol-codeine in treating post-laparotomy pain, with fewer side-effects 25. In addition, NSAIDs were superior to codeine for analgesia after second trimester abortion 26 and for the treatment of pain following vaginal delivery 27. Pharmacogenetic testing (the AmpliChip CYP450 Test) 28 is a promising tool for customising codeine treatment, but the costs associated with testing (£365–790 (€437–946; $600–1300) and the availability of equally efficacious NSAID alternatives do not justify its routine use in clinical practice. With the availability of superior and more reliable alternatives such as NSAIDs, the continued use of the less efficacious, and more unpredictable, codeine for postpartum analgesia should be questioned. Most opioids, including codeine, are readily secreted into breast milk. The actual effect of the drug on an infant is determined by a variety of factors: milk/plasma drug concentration ratio; volume of milk ingested; age of the infant; and the presence of active metabolites in the mother and infant 29. Maternal administration of codeine results in infant plasma morphine concentrations that are only 2-4% of the toxic level in infants 30 and therefore it is probably safe to use in the immediate postpartum period in the majority of women. However, recent studies suggest that maternal codeine may cause depression of the infant central nervous system (CNS) in a dose-dependent manner; the reported dose ranges that caused symptomatic CNS depression in neonates were 1.4-1.6 mg.kg−1.day−1, with a 71% concordance between maternal and neonatal CNS depression 31. Doses of less than 1 mg.kg−1.day−1 were not associated with adverse neonatal effects in these studies. However, patient-reported CNS depression is subjective, unreliable, and may not necessarily predict true complications. In fact, a large population-based retrospective cohort study confirmed that postpartum use of codeine was not associated with infant death or hospitalisation 32. Currently, serious complications such as respiratory depression and apnoea are limited to isolated case reports, especially in breastfeeding mothers who are ultra-rapid metabolisers 33. The probable cause of fatal respiratory depression in these cases is thought to be an acute increase in plasma levels of morphine and its subsequent transfer into breast milk, coupled with the inability of the neonate to handle the morphine overload 34. Neonatal abstinence syndrome can occur following codeine therapy even in the absence of maternal addiction or dependence 35. A prudent approach, therefore, would be to avoid codeine in high-risk ethnic groups such as Ethiopians, limit the dose to 1 mg.kg−1.day−1 in other mothers, move to NSAIDs as soon as possible, and closely monitor the mother-infant pair for evidence of CNS depression. A better option, perhaps, is to replace the routine use of codeine with NSAIDs in this population. This conclusion is fully supported by the Academy of Breastfeeding Medicine (ABM), which recommends non-opioid analgesics as first-choice therapy in breastfeeding mothers 36. Tonsillectomy is the most common intermediate-sized operation undertaken in children that is associated with substantial postoperative pain lasting at least seven days, requiring ‘step-up’ analgesia (i.e. additional ‘as required’ analgesia when paracetamol and NSAIDs alone are insufficient). Tonsillectomy is decreasing in frequency in Western countries but an increasing proportion is being performed for obstructive sleep apnoea rather than recurrent tonsillitis; Erickson et al. 37 found that upper airway obstruction as an indication for tonsillectomy/adenoidectomy increased in the USA from 12% of patients in 1970 to 77% in 2005. A significant proportion of children with obstructive sleep apnoea are acutely sensitive to the respiratory depressant effects of opioids 38 and caution should be exercised if these patients require other intermediate-sized procedures aside from tonsillectomy/adenoidectomy. The tragic death described at the start of this editorial 1 can probably be attributed to a ‘perfect storm’ combination of obstructive sleep apnoea, a painful operation, prescription of regular codeine postoperatively and ultra-rapid metabolism. In the face of the MHRA's pronouncement on the use of codeine in children, what are the alternatives? Dihydrocodeine is much less potent than morphine and most preparations do not have a license in children under the age of 4 years; in addition, a small proportion undergoes metabolism to an active metabolite (dihydromorphone) and it is formulated in alcohol. Tramadol is an effective analgesic but it is metabolised by the 2D6 enzyme system to an active metabolite with a high affinity for the μ opioid receptors, thus increasing the likelihood of nausea and vomiting. In addition, there is an important pharmacodynamic interaction with ondansetron, since the analgesic effect of tramadol is partly due to increased release of serotonin, and ondansetron is a serotonin antagonist 39, 40. Potential also exists for a pharmacokinetic interaction because ondansetron is, in part, also metabolised by the 2D6 enzyme system. The concurrent use of the two drugs results in a mutual reduction in effect; tramadol is a less potent analgesic and ondansetron is a less effective anti-emetic. Oral morphine is an attractive alternative in that it has reasonable bioavailability, is not a pro-drug and its metabolism does not involve the 2D6 enzyme system. However, it does have active metabolites that may accumulate with repeated dosage, and it is a controlled drug (Schedule 5 under the Misuse of Drugs Regulations 2001). Possible strategies regarding postoperative analgesia for children having intermediate-sized operations are shown in Table 1. At one of our institutions, codeine is no longer available for children and has been replaced by dihydrocodeine in the formulary. It has also been abandoned at the Hospital for Sick Children in Toronto, where it has been replaced by oral morphine 44. On the one hand, codeine is not an ideal analgesic agent, it is ineffective in a number of patients and it may actually cause death in a very small number of patients. However, on the other hand, it does have a long track record of usage, it is safe and effective in the majority of patients, it is cheap and palatable, and there is no obviously superior alternative. There have been some interesting developments, for example in the use of intranasal analgesics such as diamorphine 45 and oral tapentadol, a centrally acting analgesic with a dual mode of action (similar to levorphanol). Its potency is somewhere between tramadol and morphine but, although licensed in the UK for adults, it currently has no license for use in pregnancy, breastfeeding mothers or children. Currently, there is uncertainty as to how best to fill the so-called ‘analgesic vacuum’. All stakeholders have a duty to safeguard obstetric and paediatric patients and to provide effective analgesia postoperatively. We do not know the safest and most effective analgesic regimen to use after obstetric procedures such as caesarean section and in children following tonsillectomy. Research is urgently needed in this area. In the meantime, we require a region-based approach with specific regimens taken up by tertiary centres and adopted by their associated networks. In addition, aspects beyond a focus on pharmacology are required, including education of patients and parents and easier access to effective alternatives, especially for pain relief at home after intermediate surgery. We thank Dr M. Tremlett for his helpful advice during the preparation of this editorial. CRB contributed to the RCPCH/APA/RCoA joint statement. No other external funding or competing interests declared.