Pro–con debate: is codeine a drug that still has a useful role in pediatric practice?

Codeine is a drug that remains in world-wide clinical usage despite the availability of other opioids that are far better understood in terms of pharmacodynamics, pharmacokinetics, and side effects. Its continuing usage in clinical practice highlights a paradoxical inconsistency in the prescribing habits of codeine in the pediatric and pediatric anesthetic world. On one hand, codeine has been used extensively for many years as step-down drug from controlled opioids such as morphine, and as such continues to be regarded as ‘safe and effective’ by physicians and the general public. In most countries, it can be bought by the general public without a doctor’s prescription, which represents a tacit endorsement of its reliability. On the other hand, there is a body of evidence indicating highly variable unpredictable analgesia and side effects compared to other agents, with potential extremes of response varying from lack of effect to life-threatening complications. At the heart of the debate is the question of what are the active components of codeine and its metabolites. If codeine was being developed currently, a significant amount of new scientific clinical data would be required before it would become licensed. In trying to come to a balanced view, it is useful to try to look at both sides of the argument: from that of an experienced clinician who uses the drug regularly and from a clinical pharmacologist who can put the evidence from this stance into context. Codeine is commonly used to provide postoperative pain relief in children. It is listed by the World Health Organization (1) as the second step on the analgesic ladder for the treatment of cancer pain. It is added in if simple oral analgesics [Step 1 – paracetamol and nonsteroidal anti-inflammatory agents (NSAIDs)] have been inadequate, and before introducing potent opioids (Step 3). Its role in postoperative pain in children is very similar. It is given for postoperative analgesia after many common children’s operations as part of a multimodal approach, when paracetamol and NSAIDs are insufficient to control pain. In addition, with the extension of day-stay surgery and reduced in-patient stay, this opioid is particularly useful in providing analgesia as part of an analgesic cocktail following discharge home. Codeine is an effective analgesic for moderately severe pain available for easy oral administration both in hospital and at home. It is available in a formulation (syrup) which is easy to give by both nurses and parents and is easy to prescribe in that it is not regarded as a controlled substance. The effective dose has been established by long-term clinical usage and is easy to remember It should not be the drug of choice for acute severe pain (e.g: postcraniotomy) where morphine remains the drug of choice, but can be used effectively for moderate pain. Specific problems in a minority of patients (newborns) should not prohibit its use for the majority. No safer, fully investigated, more effective oral agents for moderate pain exist. Recent work suggests we are still struggling to provide effective pain relief for many children after surgery. A recent snap shot study of the level of pain experienced by hospitalized children in the Hospital for Sick Children, Toronto showed 64% of children had experienced moderate or severe pain in the 24 h before interview, with 23% in significant pain actually at the time of interview (2). There are no snap shot studies showing the typical levels of pain experienced by surgical children after discharge home. As pediatric anesthetists, we have a moral and ethical duty to provide effective pain relief for children after surgery, but in general, the available analgesic options have not changed significantly in the last decade. These include potent opioid analgesia such as morphine, weak opioids such as codeine, nonsteroidal anti-inflammatory drugs (NSAIDs), paracetamol, and local anesthetics. Both ketamine and clonidine have uses in specific situations endorsed by enthusiastic individuals, but they have failed to enter the mainstream of commonly used analgesic agents. With our limited armamentarium, it is crucially important that in our quest for scientific evidence-based purity, we do not discard a useful, safe, and effective agent when more current strategies are still struggling to provide reliable pain relief to our pediatric population. Codeine has been used for many years as an analgesic, but remains controversial. A recent study concluded that ‘The myth that codeine is a potent and efficacious analgesic must be exposed so that clinicians make more rational choices when managing and treating pain (3).’ Similarly, Williams and colleagues, in a review of codeine in paediatric medicine, concluded that ‘The popularity of codeine in children is not supported by convincing data of its efficacy or suitability, despite its apparent good safety record (4).’ If it is clearly so inappropriate to continue using agents that are ineffective, then why is codeine still being used? From the perspective of the practicing clinical anesthetist, there are three important questions to consider: Is codeine an effective analgesic? Is it safe? Are there better alternatives? In answering these questions, the underlying pharmacokinetics is relevant, but only in so far as it helps us to answer these questions. In the United Kingdom, no central records are kept of the number of prescriptions for codeine preparations issued by hospital pharmacies per year. However, discussions with colleagues across the United Kingdom suggest that codeine is still a commonly used analgesic and is felt to be a valuable agent. Nevertheless, clinical dogma is not enough and begs the question: ‘Is there a body of clinical evidence to support or refute its use?’ This is difficult with codeine as, along with many other established agents, there are no large randomized controlled trials of codeine in discreet, age-specific pediatric populations, and clinical situations. It is therefore necessary to consider adult studies to decide whether codeine is a useful and effective as an analgesic. While this is not ideal given the known differences in pharmacokinetics and pharmacodynamics that occur with maturation, the available data from other opioids would indicate that the major maturational changes in drug action occur within the first year of life (5, 6). Codeine is most commonly used in combination with paracetamol. A Cochrane Database review in 1998 compared the effectiveness of a range of single doses of paracetamol alone and in combination with codeine for moderate to severe pain (7). They calculated the number needed to treat (NNT) as a measure of analgesic efficacy. NNT was defined as the number of patients needed to receive a treatment for one patient to achieve at least 50% pain relief over 4–6 h compared to placebo. For postoperative pain, paracetamol 1000 mg had a NNT of 4.6 (95% confidence interval 3.8–5.4). The NNT for paracetamol 1000 mg plus codeine 60 mg was 1.9 (95% confidence interval 1.5–2.6). This figure was derived from data on only 127 patients in two trials and so is unreliable. The evidence was reassessed in 2001 by the Pain Research Unit of the University of Oxford using slightly broader but still stringent study inclusion criteria (8). They were able to identify six additional trials of paracetamol 1000 mg plus codeine 60 mg. With greater patient numbers, the number NNT for these therapeutic doses of paracetamol and codeine was 2.2 (confidence interval 1.7–2.9). The addition of codeine to paracetamol would appear to provide improved analgesia as a single dose in adults. These results may not reflect the true value of codeine as an analgesic. Codeine is not given as a single dose with a single dose of paracetamol but on a regular basis. A clinical study using extraction of the third molar tooth as a standardized pain experience showed pronounced improvement in pain relief with the second dose compared to the first dose of codeine 60 mg orally (9). For the experienced clinical anesthetist, the postoperative visit is as important as the preoperative assessment. It highlights the effectiveness of the initial part of the plan for postoperative analgesia and any associated morbidity. The outcome once discharged from hospital is unseen by anesthetists but falls to primary care. Wolf (10), in an editorial, on day case surgery almost a decade ago, highlighted the situation of initially good pain relief in hospital not continuing once discharged home. Any child or parent will testify that pain does not stop as the child leaves hospital. Tonsillectomy remains one of the commonest operations undertaken on children. It causes considerable pain lasting more than 7 days. The pattern is of intense or moderately intense pain for the first 3 days, followed by a gradual decline over the next 4 days (11). Pain assessment and management falls to parents or guardians. There are no published studies on pain relief in children at home with different analgesic regimes with or without codeine. Three years ago, the author (MT) was prepared to consider the view that codeine was not helpful as an analgesic and prospectively audited 60 tonsillectomies postoperatively. All were 23 h discharges and were advised to administer therapeutic doses of both paracetamol and ibuprofen on a regular basis for a week. They were asked to measure usual and worst levels of pain experienced on a daily basis using the Wong and Baker Faces scales and were telephoned days 3 and 5 postoperatively. Pain scores for worst pain experienced in previous 24 h were a median of 4 of 5 on Day 2 for both parent and child assessments remaining at 3 on Days 4 and 5. Fourteen of sixteen (23%) children had consulted their general practitioner by Day 5 because of waking in the night (commonly between 2 and 4 AM) with significant pain. Since then we have added codeine elixir to be given as required for breakthrough pain. Ongoing prospective audit is under way with only 1 of the first 29 (3%) children seeing general practitioners with inadequate pain relief by Day 5 postoperatively. This personal experience is confirmed by Ewah (12). She prospectively followed up 100 Day Case tonsillectomies for 3 days after operation. All received ibuprofen, paracetamol, and codeine regularly. None had visited general practitioners by Day 3 (personal communication), and low median pain scores were reported. Experience from two separate anesthetists auditing the effectiveness of their routine clinical practice attests to the effectiveness of codeine for pain relief at home following intermediate surgery. There are two reasons why this information does not exist in the scientific literature. First, there is no financial drive from the pharmaceutical industry to investigate this established and cheap agent. Second, the increasingly regulation around clinical investigations has reduced the ability of practicing clinical anesthetists from undertaking simple patient studies. Codeine, like all drugs, is not free of problems. It should not to be given intravenously as it may cause profound hypotension (possibly from histamine release) and grand mal convulsions. In addition, there are a number of specific patient groups where its use is associated with significant risk of complications. There is a recent, well-documented case of death of an infant (1 month old), who received combination medications including codeine as an anti-tussive and developed profound respiratory depression at home. The total codeine dose was 1.26 mg·kg−1 in a 6-h period (13). The use of codeine is also now not advisable in breastfeeding mothers. A case is reported of a breastfed baby dying from respiratory depression (14). Genotype analysis revealed that the mother to have a gene duplication (three functional CYP 2D6 alleles) and thus to be an ultra-rapid metabolizer (UM) of codeine to morphine. Analysis revealed high levels of morphine in expressed breast milk. As clinicians, we use codeine orally on a regular basis as part of a multimodal approach to pain management for a number of days at home. The children are over the age of 1 year and usually otherwise well. The earlier case reports are important but do not contraindicate our use of codeine for the majority. Despite its widespread clinical use, there has been only one recent case report of apnea in a child after oral codeine (15). It was attributed to the child being an UM of codeine to morphine. Subsequent correspondence casts very considerable doubt on the claims made in the case report. Therefore, use of codeine for many years and in many patients suggests a good safety record. If one accepts that codeine is an analgesic, then the arguments against its use include concerns relating to: Pharmacokinetics (genetic polymorphism). Adverse effects (side effects) The first is that the analgesic effect of codeine is because of its conversion to morphine by an enzyme showing substantial genetic polymorphism. The enzyme system involved is the cytochrome P450 2D6 enzyme system. Nine percent of UK patients are estimated to have low levels of the enzyme and are described as poor metabolizers (PM) (4). They form low levels of morphine and therefore would be expected to get little or no analgesia from morphine. A number of individuals show enzyme duplication and are described as UMs. They may produce abnormally high levels of morphine and are therefore at risk of profound respiratory depression. Clinically, we do not know the genotype of any individual postoperative child, and thus, we are at risk of under or overdosing the child. The case that codeine works by conversion to morphine appears to be unproven. Normally, very little of the total dose of codeine is actually metabolized by CYP 2D6 to morphine, the range being between 0.5% and 2–3%. Eighty percent of codeine is directly metabolized by the phase 2 uridine diphosphate glucuronosyl transferase (UGT) enzymes of the liver to form codeine-6-glucuronide (C6G). Veer et al. (16) have postulated that C6G is responsible for most of the analgesic action of codeine by activity at mu receptors. Interested readers are referred to Armstrong and Cozza (17) for a fuller discussion. Codeine has been regarded by some as having all the disadvantages of morphine without the advantages, meaning high levels of side effects without analgesia. Codeine certainly causes constipation. This is well recognized and is readily treated. There is however, little evidence of high incidences of other side effects. The Cochrane review (7) found that the addition of codeine to paracetamol did not result in a significant change in the incidence of nausea and vomiting. However, they did find a small increase in the incidence of two side effects (drowsiness/somnolence and dizziness) with the addition of codeine. If it is true that codeine is little more than an unreliable pro-drug of morphine, then from a pharmacologists perspective should we be sending children home of low dose morphine rather than codeine? This is fraught with practical difficulties. In the United Kingdom, codeine is classified as a Class B drug under the terms of the Misuse of Drugs Act 1971, and only a Schedule 5 drug (if not in injectable form) under The Misuse of Drugs Act 2001. The Schedule 5 status exempts oral codeine from almost all of the regulations surrounding controlled drugs. Morphine is a Class A, Schedule 2 drug. This demands very specific requirements when prescribing the drug, including that the drug is kept in hospital in a locked receptacle (cupboard) and requires a Register of drug administration to be kept in line with the Act. These regulations would make routine dispensing of low dose oral morphine as ‘Take home analgesia’ from Pediatric Surgical Day Case Units almost impossible. Are any of the newer opioid analogs better alternatives for providing acute postoperative pain relief than codeine? Buprenorphine sublingually is not licensed for children under 6 years of age. Tramadol is not licensed under the age of 12, and Oxycodone has no pediatric licence at all. There are very little data available to say if any of these agents are realistic alternatives to codeine. Codeine or methylmorphine is an opioid used for its analgesic, anti-tussive and anti-diarrheal properties. Codeine is an alkaloid found in raw opium in concentrations commonly ranging from 0.2% to 0.8%. Opium is full of other powerful active analgesic compounds, and codeine’s analgesic action is mediated through its metabolite, morphine. Codeine was first isolated from opium in 1832, but proprietary codeine is nowadays synthesized from morphine through the process of O-methylation. Codeine use has decreased in pediatric anesthesia from its heyday 20 years ago when intramuscular (IM) injection provided postoperative analgesia after tonsillectomy and when it was routine medication in neurosurgical units. Pleas for re-evaluation of its use have been made (18). The continued use of this minor opium alkaloid for pediatric analgesia remains baffling. Reasons may include: familiarity with the name, licensing laws that restrict appropriate use of better analgesic drugs and ignorance. Codeine is a pro-drug of morphine, so why not use morphine? Analgesia may be inadequate after oral or IM codeine. Titration of intravenous morphine by patient or nurse provides superior analgesia (19). Codeine costs six times that of similar morphine formulations. Protocols for morphine by rectal, oral, subcutaneous, and intravenous are well documented in children. This is not the case for codeine. Developmental pharmacokinetics of codeine is not fully elucidated. Pharmacogenomic phenotype maturation is not reported. Bedside gene probes that might help predict a child’s drug sensitivity are not yet in common use. Pediatric formulations for combined preparations that have reasonable effectiveness are unavailable. Subdivision of currently available tablets (e.g. paracetamol 1000 mg + codeine 60 mg) is a crude process. Each administration route for codeine is associated with adverse effects. IM injection is now uncommonly used in children because it hurts. Intravenous codeine is contraindicated, because it can cause hypotension, facial swelling, pulmonary edema, and convulsions. Oral administration is associated with GI upset. Codeine is a drug familiar to both lay and medical personnel from its use as an analgesic, anti-tussive, and anti-diarrhea agent. Regulations governing availability vary from country to country. In some countries, it is available without prescription as combination preparations from licensed pharmacists in doses varying from 8 to 15 mg·tablet−1. These combined preparations often comprise mixtures of aspirin, paracetamol, caffeine, doxylamine, or ibuprofen. Formulations that included phenacetin gave codeine notoriety when patients taking this combined medicant suffered analgesic nephropathy (20). Codeine is currently receiving increased attention as a drug of abuse (21-23), and neonatal abstinence syndrome and even cerebral infarction following maternal codeine use during pregnancy has been reported (21-25). Regulations govern the prescription and dispensing of morphine. Morphine is a favored opioid analgesic in children, but the ability of hospital nursing staff to rapidly administer morphine can be hampered by regulations. The drug is often in a locked cupboard, the keys to which may be held by one individual and two signatures may be necessary for dispensing. Less restrictive regulations surrounding codeine make it an easier opioid choice. The popularity of tramadol, another drug acting on the opioid μ-receptors, can be partly explained by similar less restrictive regulations governing use. Codeine has lower affinity for μ-receptors than morphine and reduced effectiveness. Much of its analgesic effect may be mediated through its metabolite, morphine. Another metabolite, C6G, is speculated to be responsible for some analgesia, but human data are scant (26). An analgesic ceiling effect for codeine is apparent with increasing dose only causing increased adverse effects. This has been ascribed to a dual effect: analgesia being mediated through morphine but adverse effects predominantly because of codeine receptor occupancy (27). There is little evidence for the broad belief that codeine causes less adverse effects, such as sedation and respiratory depression, compared with other opioids (4). Glucuronidation (UGTB27) is the major metabolic clearance (CL) pathway, but minor pathways are N-demethylation (cytochrome P450 enzyme CYP3A4) to norcodeine (10–20%) and O-demethylation (CYP2D6) to morphine (5–15%). Codeine is also excreted unchanged in the urine (5–15%)(4). N-demethylation is comparable to that in adults in the perinatal period, while the other metabolic pathways develop beyond the newborn period. Although maturation of codeine CL with age has not been reported, it is expected to follow that of morphine, because glucuronidation is the major pathway, where maturation is mostly complete by 1 year of age (28). Oral codeine is rapidly absorbed with approximately 50% of the dose undergoing first pass metabolism. Peak plasma concentrations occur ∼1 h, and the plasma half-life is 3–3.5 h in adults (29). IM absorption speed is similar to that by the rectal route, although the latter is associated with variable absorption (30). The pharmacokinetics of codeine is poorly described in children despite use over decades. A volume of distribution (V) of 3.6 l·kg−1 and a CL of 0.85 l·h−1 have been described in adults, but there are few data detailing pediatric pharmacokinetic developmental changes. The neonatal half-life is longer because of immature CL (e.g. 4.5 h), while that of an infant is shorter (e.g. 2.6 h) (31) attributable to size factors (32). It would be useful to understand maturation of the CYP2D6 enzyme, because this enzyme is central to analgesic effect. In vitro studies suggest that CYP2D6 enzyme activity develops over the first month of life to reach 60% of adult activity by 28 days (33, 34). There are no in vivo data describing or quantitating this maturation in neonates. However, tramadol is another drug with metabolism predominantly by CYP2D6. Tramadol CL increases from 25 weeks postconception age (PCA) where CL is 5.52 l·h−1 per 70 kg and reaches 84% of the mature value by 44 weeks PCA (35). These data are consistent with in vitro results, but the impact of CYP2D6 polymorphism on the variability in pharmacokinetics remains to be established. Tramadol CL maturation data suggest that morphine formation from codeine will be very low in premature neonates, irrespective of the CYP2D6 polymorphism with subsequent rapid maturation. The slope of this increase depends on the CYP2D6 activity score (36). We might expect almost no analgesic effect in premature neonates, and that there may be effect beyond 44 weeks, but this effect is dependent CYP2D6 polymorphisms. More than 60 alleles in the CYP2D6 gene have been identified (37). Depending on the number of dysfunctional, less active, or normally active alleles an individual possesses, phenotype can be classified as poor, intermediate, or extensive metabolizer. Individuals with duplicated active CYP2D6 genes are classified as ultra-extensive metabolizers (38, 39); 7–10% of Caucasians are slow metabolizers. This percentage varies with ethnicity (39-41). For example, 2% of Asians and 1% of Arabs have poorly functional CYP2D6, and codeine should be less effective for analgesia in these patients. Codeine has minimal analgesic effect in the PM (42), but adverse effects (e.g. sedation, pruritus or nausea) persist (41). Gastrointestinal motility is affected only in extensive metabolizers; suggestive of a morphine-dependent mechanism of action for this (43). Factors that might decrease conversion of codeine to morphine influence the amount of morphine available from codeine. Immaturity of CYP2D6 in neonates renders the drug impractical in this age group. Selective serotonin reuptake inhibitors such as fluoxetine and citalopram inhibit CYP2D6 activity. Rifampicin and dexamethasone induce expression of CYP450 isozymes and increase the rate of metabolism. Concomitant medications competing for the CYP2D6 enzyme pathway (e.g. amiodarone, cimetidine, celecoxib, hydroxyurea, metoclopramide) may reduce CL (4, 44). Codeine is less potent than morphine (potency ratio 1 : 10). Consequently, 60 mg of codeine only has morphine equivalence of 6 mg. Oral codeine is rapidly absorbed with approximately 50% of the dose undergoing first pass metabolism. Codeine has higher oral/parenteral potency ratio (2 : 3), because hepatic conversion to morphine is necessary for codeine’s analgesic activity, while the first pass effect after oral morphine reduces bioavailability of morphine (approximately 30%) (45). The NNT to produce 50% pain relief for 4–6 h in moderate to severe pain is a useful measure of relative effect. The combination paracetamol 1000 mg/codeine 60 mg provides better analgesia (NNT 1.9) than codeine 60 mg (NNT 17) or paracetamol 1000 mg (NNT 4.6), illustrating the synergistic effect of analgesic combinations (18). Codeine continues to be used for postoperative analgesia after neurosurgical procedures (30, 46). This practice evolved from the belief that codeine causes less sedation and respiratory depression than morphine, despite data showing morphine is more potent, longer lasting and possesses a similar safety profile (47). Such data indicate that postoperative pain would be more appropriately treated with morphine than with codeine, although results also suggest that severe pain is not a persistent problem, because only two doses of morphine were required postoperatively (47). Consequently, codeine 30–60 mg IM provides adequate pain relief in 90% of the patients with postoperative adult craniotomy (48). IM injections are now rarely used in pediatric anesthesia practice, and it would seem sensible to initiate analgesic treatment with a simple analgesic (e.g. paracetamol) and to titrate a small dose of intravenous morphine if required. Codeine continues to be used for adenotonsillectomy pain where it is associated with less postoperative vomiting than morphine while providing comparable postoperative analgesia (49) This observation is attributable, in part, to the synergistic effect reported when codeine is combined with paracetamol (acetaminophen). In addition, the peak morphine concentration is higher after an intravenous bolus of morphine than that associated with an equivalent codeine dose given orally. Vomiting is more frequent after morphine 0.15 mg·kg−1 than after codeine 1.5 mg·kg−1. The overall probability of vomiting after morphine 0.1 mg·kg−1 was 50% (50). IM injections are infrequently used in children, and titration of smaller morphine doses is associated with less vomiting. Current management of tonsillectomy pain is multipronged in approach often incorporating local anesthesia, paracetamol, NSAIDs, tramadol, and steroids. Opioids can be avoided or at least be restricted to shorter acting agents such as fentanyl with morphine constrained to rescue medication. While the WHO recommends the use of codeine in pediatric palliative care (1), there are no data to promote this drug in preference to other opioids. Adverse drug reactions associated with the use of codeine are those common to opioids: euphoria, itching, nausea, vomiting, drowsiness, dry mouth, meiosis, orthostatic hypotension, urinary retention, and constipation. However, it appears that patients develop tolerance to many of these effects with prolonged use. Respiratory depression is the major adverse effect that concerns pediatric anesthetists. This complication relates, in part, to pharmacogenetics. Recently, a newborn died from morphine poisoning when his mother used codeine while breastfeeding. The mother, an UM, produced much more morphine when taking codeine than most people do (51). A 29-month-old previously healthy child experienced apnea resulting in brain injury following a dose of codeine 2 days after an uneventful anesthetic for tonsillectomy. A genetic polymorphism leading to ultra-rapid metabolism of codeine into morphine resulting in narcosis and apnea was proposed (52). These instances question the value of this drug in pain management. Patients presenting an UM genotype which is particularly frequent in Africans and Arabians are at increased risk to suffer from codeine-induced respiratory depression (53). The decline in codeine from pediatric practices matched the institution and development of dedicated pediatric pain services and implementation of analgesic protocols (54, 55). Drugs such as morphine and paracetamol where the pharmacokinetics (56-59) and pharmacodynamics (60-62) are understood have become routine. Morphine protocols for intermittent dosing, PCA, and infusion have become standard practice (54, 58, 63). Even the NSAIDs, so long accused of causing postoperative tonsillectomy bleeding are losing their evil reputation (64, 65). Relatively, new drugs such as tramadol have had dramatic increases in popularity over codeine. 1. The natural history of older drugs is that they are usually replaced by new agents that have improved therapeutic profiles and risk benefit ratios. Examples would include induction agents (propofol) local anesthetics (ropivacaine and l-Bupivacaine) and anesthetic vapors (sevoflurane and desflurane). Why then do you think that despite its acknowledged flaws, codeine continues to be so widely used. MT: I think codeine continues to be used for a number of reasons. Firstly, it appears to work when used in combination with other relatively weak oral analgesics such as paracetamol. It is available in an acceptable oral formulation unencumbered by legal restraints on prescribing, and no better alternative has become available. BA: Morphine is an extremely old opioid that has been used over millenniums in various guises. Codeine, like papaveretum that has now almost disappeared from practice, is simply another morphine preparation. Unlike papaveretum, it is commonly used out of hospital, because it is readily available as an oral preparation, and it is synergistic with paracetamol. It may be preferred over the NSAIDs for postoperative tonsillectomy, despite data attesting to reasonable safety and reduced postoperative nausea and vomiting with NSAIDs. Fear of paracetamol hepatotoxicity and lack of understanding of the pharmacokinetics have resulted in reduced administration doses with concomitant reduced effectiveness, necessitating another drug to fill the analgesic vacuum. Dr Tremlett has produced some nice data showing the effectiveness of adding codeine to our currently available analgesic mix. In addition, codeine’s availability over the counter ensures continued use. 2. Both authors have agreed that there are issues with being able to provide effective analgesia at home in children without using regulated drugs such as morphine that can be associated with administrative/legal difficulties. What are the feasible alternatives to codeine in providing reliable analgesia and what would you recommend from your own clinical practice? MT: This is an important question. We would not be having this debate if we had all the answers about codeine in pediatric practice. Firstly, it is important to recognize why and how we are using codeine at present and then consider alternatives in that role. The place of codeine for me is to provide analgesia for moderately severe pain where we know the combination of paracetamol and/or NSAIDs to be inadequate. The only real contender in place of codeine in my mind is Tramadol. BA: Paracetamol administered in conjunction with NSAIDs is a powerful combination, and the vast majority of patients are sent home with this combination. The correct mix of these drugs remains poorly quantified and probably depends on relative mix ratios. The literature is confusing, because some studies show no advantage to adding paracetamol to NSAIDs, while others demonstrate additivity. Tramadol is being used increasingly in our practice, despite licensing limitations. It should be pointed out that just because a drug is not licensed for pediatric use, this does not exclude its clinical use. 3. What evidence is available to support the use of tramadol or other opioids agents in preference to codeine? MT: I have no first hand experience of the use of tramadol for the management of pain in children. The literature available is also very limited. A recent review had to conclude only that ‘the real place of tramadol in the armamentarium of postoperative analgesics in children is undefined ! (66).’ It will only replace codeine in my practice if we have good studies comparing the clinical effectiveness and side effects of tramadol vs codeine in pediatric populations. BA: Dr Tremlett points to the lack of pediatric studies investigating postoperative analgesia after discharge from hospital. Adult data suggest that the withdrawal of dextropropoxyphene from practice will have little impact (67). Tramadol is gaining increasing popularity, but the high incidence of nausea in children can be a negative adverse effect. Adult data from postoperative orthopedic pain suggest similar pain scores for tramadol–paracetamol compared to codeine–paracetamol, although the incidence of constipation and vomiting was higher in those given codeine (68). Other data suggest increased somnolence in adults given codeine over tramadol (69); although this may not necessarily be a bad thing in children. The (+)-M1 metabolite of tramadol is also formed via the genetically polymorphic CYP2D6 iso-enzyme system responsible for codeine metabolism, and individuals may be classified as extensive or PM of tramadol. Higher concentrations of the (+)-M1 metabolite and greater analgesic efficacy of tramadol are reported in extensive metabolizers with reduced nausea, vomiting, and tiredness among PM. Oral morphine is an acceptable alternative to codeine in hospital practice. Taste, however, can be problematic, and regulations restrict use out of hospital. Transdermal approaches using fentanyl and buprenorphine are gaining acceptance for palliative care, but not yet for postoperative analgesia. Opioids such as oxycodone or hydromorphone are also gaining increasing use, but have not yet reached familiarity or comprehensive review to be used by most practitioners. Dipyrone is available in only a few countries. Weak opioids (e.g. codeine and tramadol) are frequently omitted in favor of the strong opioids from pain regimens in palliative care at home in both adults and children. (70, 71). In some clinics, tramadol may be the only weak opioid used (72). 4. Like many older drugs, the safety of codeine seems to be built on years of experience, so that while no individual drug can be considered completely safe, the familiarity with therapeutic effects and side effects are well established. Given this wealth of clinical experience, and despite the case reports discussed by both authors, should we simply accept that in some clinical areas such as pediatric neurosurgery where the drug has a proven record we should simply continue with this practice? MT: I accept that there is very little in the pediatric literature to support my position that codeine is, in the main, a safe and effective analgesic in combination with other drugs for moderate postoperative pain. It is important that we do not just continue what we have always been doing but actually audit the effectiveness of our practice. In particular, it is important that we look at the pain and side effects experienced by our children immediately following discharge from hospital, review our present analgesic prescribing practices and see for ourselves whether codeine is clinically valuable or not. It is also important that we do not use codeine where other drugs are more appropriate. It is not the treatment of choice for acute severe pain. BA: I would argue that oral (but not rectal (73)) morphine also has a proven record and is an acceptable alternative after pediatric neurosurgery. Pain management is increasingly in the hands of a dedicated team rather than individual surgeons. Infrequent codeine use by staff in other disciplines will lessen familiarity. It may be safer to have a consistent approach to pediatric pain throughout a hospital rather than idiosyncratic management in one unit. Again, adult data document vomiting and retching occurred in 50% of patients after craniotomy treated with tramadol, compared with 20% with morphine, and 29% with codeine (74). While codeine is associated with less sedation, nausea, and vomiting than tramadol (75), morphine remains the gold standard. It is interesting to note that while both authors have strongly argued their case, both accept that codeine is likely to remain as a popular drug in pediatric anesthesia for the foreseeable future. Although there are reasonable theoretical and clinical arguments against codeine, the drug still has a central place in mild and moderate pain through years of familiarity. Future research to optimize dosing regimens using combinations of codeine with paracetamol and NSAID’s may further improve pain management in children with mild to moderate pain. For more severe pain, there are good arguments to use oral morphine in preference to codeine. This is straightforward in a hospital setting where controlled drugs can be dispensed from a secure source. Outside of hospital, the issues of safe, legal, and effective opioid administration become more difficult because of drug legislation. Tramadol may show promise for analgesia at home but lacks an evidence base against an active comparator. Until, other agents have been shown to be more effective in a home environment with fewer adverse effects, codeine is likely to remain in its current niche.