Codeine and Opioid Metabolism – Implications and Alternatives for Pediatric Pain Management

Abstract

Purpose of review

Use of perioperative opioids for surgical pain management of children presents clinical challenges because of concerns of serious adverse effects including life-threatening respiratory depression. This is especially true for children with history of obstructive sleep apnea. This review will explore current knowledge of clinically relevant factors and genetic polymorphisms that affect opioid metabolism and postoperative outcomes in children.

Recent findings

Within the past several years, an increasing number of case reports have illustrated clinically important respiratory depression, anoxic brain injuries and even death among children receiving appropriate weight-based dosages of codeine and other opioids for analgesia at home setting particularly following tonsillectomy. Several national and international organizations have issued advisories on use of codeine in pediatrics, based on CYP2D6 pharmacogenetics. We have discussed the pros and cons of alternatives to codeine for pain management.

Summary

Although routine preoperative genotyping to identify children at risk and personalized opioid use for pediatric perioperative pain management is still a distant reality, current known implications of CYP2D6 pharmacogenetics on codeine use shows that pharmacogenetics has the potential to guide anesthesia providers on perioperative opioid selection and dosing to maximize efficacy and safety.

Introduction

Respiratory depression is the most serious adverse effect of opioids as it could potentially lead to hypoxic brain injury and fatality (1). Opioids are responsible for fifty percent of postoperative respiratory failure events (2, 3). Unpredictable large inter-patient variations in opioid responses and narrow therapeutic indices of opioids result in a high incidence of postoperative opioid related respiratory depression (up to 41%) (4, 5), especially in children who are sensitive to opioids and differ in physiology and pharmacology from adults (6). A national survey (data from 252 institutions) of Patient/Proxy Controlled Analgesia (PCA) practices in children<6 years reported 8 deaths and 42 respiratory depressant incidents requiring naloxone over 5 years (7). A review of the LexisNexis database (1984-2010) for deaths and complications following tonsillectomy showed that 18% of death and 8.6% of injury claims (mainly anoxic brain injury) in children are opioid related (8). While various patient risk factors for in-hospital opioid induced respiratory children in children have been described, including age<1 year, obstructive sleep apnea (OSA), obesity, being underweight, prematurity, and developmental delay (9), twin studies have also revealed significant heritability (30%) for respiratory depression from opioids (10). In fact, in pediatrics, there has recently been multiple fatalities in children from the use of codeine which has turned our attention to the role of genetics, especially Cytochrome P450 family 2 subfamily D type 6 (CYP2D6) and OSA in this regard (11), (12). In this review, we will discuss briefly the role of pharmacogenetics with respect to opioid induced respiratory depression, highlighting the implications of CYP2D6 genetics on codeine safety, pros and cons of alternatives to codeine, and the role of OSA as a major risk factor for opioid induced respiratory depression.

Codeine and CYP2D6

A) Codeine Metabolism

Codeine is a weak opioid that was endorsed by the World Health Organization as the second step on the analgesic ladder for cancer pain and has been used routinely for postoperative and for breakthrough pain in chronic sufferers. It is a pro-drug with a 200-fold weaker affinity for μ-opioid receptors than morphine; while 80% of the administered drug is inactivated by glucuronidation to codeine-6-glucuronide by uridine 5′-diphosphate glucuronosyltransferase-2B7 (UGT2B7) and N-demethylation to norcodeine by CYP3A4, 5-10% of codeine undergoes O-demethylation to morphine, its active form via CYP2D6.(13) (Figure) Without O-demethylation, codeine confers a small fraction of the analgesic potency of morphine, and much of its analgesic effect is likely contributed by a metabolite, codeine 6-glucuronide.(14)

Figure.

CYP2D6 metabolic pathway of common oral opioids resulting in biologically active metabolites. Depending on CYP2D6 metabolic activity (poor, intermediate, extensive or ultrarapid metabolizing status) varying levels of respective active metabolites from codeine, tramadol, hydrocodone and oxycodone are formed resulting in clinically unpredictable inter-individual variations in responses. Oral morphine, hydromorphone and tapentadol are not affected by the CYP2D6 metabolic pathway.

B) Codeine Use in Pediatrics

Codeine with and without acetaminophen has been a commonly prescribed medicine for pain (>18 million US outpatient prescriptions in 2013)(15) as it is relatively inexpensive, widely available in both liquid and tablet form and was considered to be tolerated well. In fact, a study from 2011 reported that codeine was prescribed to >800 000 patients less than 11 years of age, more than any other opioid(16). Codeine was mostly prescribed by otolaryngologists (19.6%), dentists (13.3%), pediatricians (12.7%), and general practice/family physicians (10.1%). Hydrocodone-containing analgesics were recent rescheduled from the Drug Enforcement Administration (DEA) Schedule III to the more tightly regulated Schedule II. This left codeine coformulated with acetaminophen as the only opioid analgesic classified as a Schedule III controlled substance and cough formulations with codeine remain Schedule V under federal law(17); unlike their Schedule II counterparts, Schedule III regulations allow for verbal and facsimile prescribing to pharmacies as well as refills with the original prescription. However, reports of fatalities in children from the use of codeine (18, 19) have raised concerns about the safety of codeine use (20-22). This was emphasized by the American Academy of Pediatrics recently in a report released recently titled “Codeine: Time to Say “No”.(12)

C) CYP2D6 polymorphisms and Phenotypes

Variability in the clinical response to codeine prompted investigations into genetic variants or polymorphisms of CYP2D6. This enzyme is mapped to chromosome 22 at 22q13.1. More than 100 polymorphisms (functional and nonfunctional) of CYP2D6 have been described to date (http://www.cypalleles.ki.se/cyp2d6.htm) (23). CYP2D6*1 is the wild-type allele and is associated with normal enzyme activity and the “extensive metabolizer” (EM) phenotype. The CYP2D6 alleles *2, *33, and *35 are also considered to have near-normal activity.

The mutant alleles, *3, *4, *5, *6, and*9, confer no CYP2D6 activity (23-25) and account for more than 90% of poor metabolizer phenotypes. Variants *10, *17 and *41 have modestly reduced activity and are referred as intermediate metabolizers (IM) (25). An individual who has multiple copies of functional genes(24), would have the Ultrarapid metabolizer (UM) phenotype. Genetic testing is commonly available for common CYP2D6 variants. An activity score is assigned to each allele in the diplotype (0 for nonfunctional, 0.5 for reduced function, and 1 for each copy of a functional allele). The patient's predicted metabolizer phenotype, is defined by the sum of the two scores:

• PM has an activity score of 0

• IM has an activity score of 0.5

• EM (normal) has an activity score of 1 to 2

• UM has an activity score greater than 2

There are inter-ethnic differences in frequency of these phenotypes; while 10% of Caucasians and 30% of Hong Kong Chinese are PM(25), 1% in Denmark and Finland, 10% in Greece and Portugal and 29% in Ethiopia(26) are UM. Hence, while codeine may be less effective as an analgesic in about 2-10% of ethnic groups (27, 28), it could be a dangerous analgesic in the latter populations, as excessive doses of morphine may be rapidly produced (25).

D) Reports of Codeine fatalities

Several deaths or near deaths have been reported with “standard” doses of oral codeine in children later found to be UM of CYP2D6 (18, 29). In 2009, a fatality after codeine administration was reported in a healthy 2-year old boy given codeine two days after adenotonsillectomy. Autopsy results showed high blood concentrations of morphine (32 ng/mL) and low codeine concentrations (0.70 ng/mL); genotyping revealed functional duplication of the CYP2D6 allele.(18) This was followed by further reports in 2012, of three deaths and two cases of respiratory insufficiency in post-surgical children, who had risk factors like OSA (19, 30). In 2013, obesity, codeine toxicity and polypharmacology were implicated in the deaths of three obese children aged four to 10 years given codeine doses based on ideal body weight(31).

E) Regulations against Codeine Use in Pediatrics and Breast-feeding Mothers

These incidents ultimately led to new regulations by the World Health Organization (March 2011)(32), US Food and Drug Administration (FDA) (August 2012)(33), European Medicines Agency (EMA)(34) and Health Canada (June 2013)(35) and the UK Medicines and Healthcare Products Regulatory Agency (July 2013 updated in April 2015)(36). Restrictions were placed on use of codeine in children under the age of 12 years, after adenotonsillectomy procedures.(37) In 2013, the joint FDA advisory committee recommended amendment of the codeine label to include a “black box warning” contraindicating codeine treatment of pain and cough in all children <18 years of age and to remove codeine from the Over-the-Counter monograph.(38) The “Warnings/sections of the label were also updated.

A recent review of the adverse event reporting systems data of children who had codeine or codeine-containing products by the FDA over past 50 years showed 64 cases of severe respiratory depression and 24 deaths mostly in children younger than 12 years of age(39).

The 2014 Clinical Pharmacogenetics Implementation Consortium (CPIC) guidelines for CYP2D6 genotype and codeine therapy(40) suggest using alternative analgesics to codeine in patients who are CYP2D6 poor for better efficacy and ultrarapid metabolizers for safety reasons. In fact, several pediatric hospitals have elected to remove codeine from their formularies.

Alternatives to Codeine

A) Non-opioid drugs

Acetaminophen and nonsteroidal anti-inflammatory drugs are good alternatives for treating mild pain, as they do not have the adverse effects of respiratory depression. Increased use of non-opioids like oral and intravenous formulations of acetaminophen and non-steroidal anti-inflammatory drugs, which do not have the respiratory depressant side effects may be good alternatives to codeine in children. Ibuprofen was found to be least as effective as acetaminophen with codeine for postoperative pain control in children after tonsillectomy and facial surgery, with no increased risks of bleeding (41-43).

Dexmedetomidine (DEX) sedation is a promising sedative agent and an adjunct to anesthetic regimen especially in patients with history of obstructive sleep apnea (OSA) (44, 45). Because of the increased risks with opioids in obese patients and those with OSA, DEX as an adjunct to anesthetic regimen may potentiate opioid analgesia with minimal additional respiratory depression. In comparison with fentanyl (1 mcg/kg), intraoperative DEX (2 mcg/kg bolus followed by 0.7 mcg /kg/hour) decreased the postoperative opioid requirements and the episodes of desaturation in children with OSA following tonsillectomy and adenoidectomy (46).

B) Tramadol

Is tramadol an alternative? Tramadol is a weak opioid agonist that is metabolized primarily through hepatic N-demethylation by CYP3A4 to an inactive metabolite, and minimally through CYP2D6- mediated oxidation to O-desmethyltramadol (M1), which has a 200-fold greater affinity for μ-opioid receptors than the parent drug.(47) Tramadol exerts its analgesic activity through complementary mechanisms: activating the μ-opioid receptor by mainly M1, and weak inhibition of norepinephrine and serotonin reuptake. A prospective, double-blinded, randomized controlled trial in children undergoing tonsillectomy compared the efficacy and safety of tramadol versus codeine/acetaminophen. The authors found that tramadol achieved similar analgesia, with less potential for side effects (48). However, studies have shown that children who are CYP2D6 PM have lower plasma concentrations of the active metabolite and analgesia compared to EM, and in 2D6 UM, plasma concentrations, analgesia and side effects are greater than in those who are EM (49-51). There was also a recent case report of tramadol administration leading to respiratory depression in a child who was a CYP2D6 UM (52). In fact, the 2010 FDA drug label reads that “concomitant administration of CYP2D6 and/or CYP3A4 inhibitors, such as quinidine, fluoxetine, paroxetine and amitriptyline (CYP2D6 inhibitors), and ketoconazole and erythromycin (CYP3A4 inhibitors), may reduce metabolic clearance of tramadol increasing the risk for serious adverse events including seizures and serotonin syndrome”. Hence, its safety needs to be further investigated before widespread use.

C) Tapentadol

A similar agent, tapentadol, is a centrally acting analgesic with a dual mode of action as an agonist at the μ-opioid receptor and a weak norepinephrine reuptake inhibitor (compared to tramadol). It has no active metabolites and mainly undergoes glucuronidation (53). A report from a poison control center reported two patients - 9-month-old child who had coma and respiratory depression, and a 16-month-old girl with dyspnea, drowsiness/lethargy, pallor, and vomiting who was admitted to critical care and treated with oxygen. However pediatric data on the safety of this medicine is limited.

D) Oxycodone

Would oxycodone, a semisynthetic opioid, be a safer alternative to codeine for managing post-tonsillectomy pain at home setting? We know that 80% of oxycodone undergoes N-demethylation by CYP3A producing noroxycodone, a metabolite with weak analgesic properties; and 10% undergoes O-demethylation to oxymorphone and noroxymorphone, by CYP2D6 (54). Oxymorphone is 14 times more potent than oxycodone (55). The affinity of noroxymorphone at the opioid receptor is three- and 10-fold higher than oxycodone and noroxycodone respectively. Hence, in order of highest to lowest potency of oxycodone and metabolites, we have oxymorphone > morphine > noroxymorphone > oxycodone > noroxycodone (56). In a postoperative setting, the CYP2D6 PM were found to have decreased oxycodone metabolism, and higher analgesic consumption (57). The greatest oxymorphone/oxycodone concentration ratios occurred in those who are UM and smallest in those who are PM (58). Other adult studies in postsurgical patients and cancer patients detected clinical differences among the CYP2D6 genotypes (59, 60). CYP2D6 activity also correlated with oxycodone experimental pain assessment, with UM experiencing increased pharmacodynamic effects (61). Oxycodone overdose by wrong dosing has been reported before (62). Interestingly, CYP2D6*4 allele (PM) implicated in the death of four fatalities from oxycodone, by a retrospective analysis of 15 cases from the Milwaukee County Medical Examiner's Office using pharmacogenomics for postmortem forensic toxicology(63). However, our ongoing pediatric oxycodone pharmacokinetic and pharmacogenetic study in perioperative setting revealed that compared to PMs, IMs and EMs have higher oxymorphone concentration. Further large studies are needed to assess whether oxycodone would be safer than codeine in children with CYP2D6 UM metabolizing status and/or with significant sleep apnea. It has also been shown that like codeine, maternal oxycodone use also causes central nervous system depression in neonates, so oxycodone is unlikely to be a safer alternative in this clinical setting (64).

E) Hydrocodone

Hydrocodone is also a potential alternative for analgesia. It is about 12 times more potent at the opioid receptor than codeine(55), and about half the clearance is via CYP2D6 and CYP3A4, into an active metabolite hydromorphone, and norhydrocodone, respectively. CYP2D6 UM may thus have up to an eightfold greater plasma concentration of hydromorphone, whereas PM receive minimal analgesia (65). A common component of antitussive medications, the US FDA banned the sale of more than 200 hydrocodone products in response to occurrence of >400 cases of adverse events associated with its use between 1969 and 2005 (66).

Hence, although both hydrocodone and oxycodone undergo metabolism via CYP2D6 to active metabolites, their analgesic effects are not as variable as codeine's because the parent drugs are nor prodrugs, but themselves pharmacologically active. One case report demonstrates the complex interplay between drug-drug interactions and pharmacogenomics. A developmentally delayed five year Somalian old child died after administration of high doses of hydrocodone for ear infection. On genotyping, she was found to be a CYP2D6 PM, and had low hydromorphone blood concentrations; but co-administration of clarithromycin (a potent CYP3A4 inhibitor) and valproic acid for seizures since birth, prevented hydrocodone metabolism leading to high hydrocodone levels(67). Hence, the dose–toxicity relationship of the alternative opioids need to be further studied in the pediatric population.

F) Oral Morphine

Lastly, the use of an oral morphine elixir has been suggested by some as an alternative(68). The prescribers and pharmacists need to be vigilant when prescribing oral morphine as it is available in many concentrations (69), however, although there is extensive experience with intravenous morphine in children, there is little clinical experience and very limited comparative clinical data on safety and efficacy available for the oral formulation.

In summary, drugs such as morphine (0.2-0.5 mg/kg/dose, every four to six hours), oxycodone (0.05-0.15 mg/kg/dose, every four to six hours), hydrocodone (0.1-0.2 mg/kg/dose, every six to eight hours) on as needed basis for severe pain are likely alternatives for severe postoperative pain, but have their disadvantages. (21).

Managing post-tonsillectomy pain safely and effectively without Codeine at home

Pediatric OSA is associated with high incidences of adverse outcomes with the use of opioid analgesics for post-tonsillectomy pain management. In addition to codeine, potentially other oral opioids metabolized by the CYP2D6 pathway such as tramadol, hydrocodone and even oxycodone, cannot be considered safe analgesics without appropriate precautions to manage post-tonsillectomy pain at unmonitored home setting especially in young children with sleep apnea. (64)

Though preoperative CYP2D6 genotyping study in children undergoing tonsillectomy is an option, it is not widely available and third party payers do not readily reimburse for the genetic tests despite the CPIC guidelines (40) and other evidences, we observed significantly more adverse effects with codeine at home even when it was administered on as needed basis (Unpublished Data: Sadhasivam, MD 2012). Preoperative CYP2D6 testing before prescribing around the clock codeine, hydrocodone, tramadol and oxycodone at home setting would be preferable as all are at least partially metabolized by CYP2D6 pathway (Figure), especially in young children with OSA and other respiratory comorbidities. Alternatively, using other analgesics (i.e. non-opioids and possibly oral morphine with appropriate doses) will lead to safer outcomes following tonsillectomy in young children.

Cincinnati Children's Hospital Medical Center Pain Management Protocol after tonsillectomy

An alternative and less expensive approach to routine preoperative CYP2D6 genotyping would be to use safer and non-opioid analgesics in young children.(70) In our institution, we changed our tonsillectomy practice (in children < 6 years) even before the 2013 FDA's warning based on many reported and unreported deaths related to opioid pharmacogenetics (71, 72) and based on our prospective pharmacogenetic research in children undergoing tonsillectomy and receiving morphine in the hospital and codeine at home. To maximize pain relief and safety, we have successfully used opioid sparing pain management at home setting following pediatric tonsillectomy. Our current practice is summarized in Table 1. We avoid all oral opioids at home setting on the day of tonsillectomy because of reports of high incidences of respiratory depression and deaths in the evening/night after tonsillectomy (69, 73). Despite our high annual volume of tonsillectomy, we have not seen any increase in incidence of inadequate pain control, postoperative bleeding or serious life threatening complications with the above pain management regimen in the last 3 years.

Table 1. Our current practice of opioid sparing pain management at home setting following pediatric tonsillectomy.

Cincinnati Children's Hospital Medical Center's Standardized Post-Tonsillectomy Pain Management Protocol
Children < 6 years	No Codeine or other opioids Acetaminophen Around the clock Dexamethasone Day of Surgery, Postoperative day #1 & 2 Ibuprofen from Postoperative day #1
Children > 6 years	Acetaminophen Around the clock Dexamethasone Day of Surgery, Postoperative day #1 & 2 Ibuprofen from Postoperative day #1 Oxycodone q4h prn from Postoperative day #1

Abbreviations: Prn = as needed basis

Future directions of research

Currently, robust evidence to change clinical practice based on underlying genetic risk factors, access to routine preoperative genotyping, affordability and payer coverage for genetic testing are limited. As compelling evidence for personalization of perioperative care based on genetic risk factors (e.g. CYP2D6 and codeine related deaths) increases, there will be better adaptability of routine preoperative genotyping and coverage of such services by third party payers. For example, many third-party payers are covering perioperative CYP2D6 genotyping for prescription of oral opioids in our pediatric institution. In future, pharmacogenetic studies also need to be complimented by epigenetic, proteomic, transcriptomic, and metabolomic information to gain additional knowledge and insight to improving personalized care as these factors may influence clinical outcome measures. Another need of pharmacogenomics research, especially in pediatrics, is genetic counseling. Genetic counselors help patients and their families understand and adapt to the medical, psychological, and familial implications of genetic contributions to clinical outcomes. As we transition from single-gene testing and genetic counseling to a full genomic medicine approach, clinical implications will get more complex.(74)

Clinical Translation: Bench to Bedside

As we routinely use the Global Positioning System to navigate maps and roads, in the future, it is anticipated that we will use a Genomic Prescribing System (GPS) to proactively identify underlying genetic risks and guide personalized care.(75) Proactive identification of patients at risk of adverse perioperative outcomes is an important first step in guiding personalized interventions with preferably Electronic Health Record (EHR) implemented clinical decision support integrating genetic risk factors and their implications for clinical interventions.

To implement pharmacogenomic based clinical decision support, there is a need for more robust study designs, independent validations, larger populations and robust statistical approaches.(76, 77) To realize the promise of personalized medicine to perioperative care, we need better evidences in terms of validating clinical association studies engaging physicians, patients, pharmaceutical industry, healthcare payers and policy makers. Others have taken a more active stance by identifying the 2D6 isoforms in their patients. For example, pharmacogenetic data gathered under the “Pharmacogenetics for kids” trial was incorporated into clinical decision support tools for the prescription of opioids in children with sickle cell disease (78).

Conclusion

Current known implications of CYP2D6 pharmacogenetics on codeine use shows that pharmacogenetics has the potential to guide anesthesiologists on perioperative opioid selection and dosing to maximize efficacy and safety. The present consensus seems to be that avoiding codeine whenever possible is the safest strategy. Currently, personalized opioid selection and dosing for perioperative pain management is still a long way. More studies are needed to improve genotype-based personalized perioperative care in children.

Key points.

• A multimodal opioid-sparing analgesia strategy reduces the need for perioperative opioid use, and improves analgesia and reduces serious risks associated with opioids in children, especially at unmonitored home setting.

• It is high time to avoid codeine and possibly other stronger opioids metabolized via CYP2D6 pathway, especially in children with significant co-morbidities such as OSA or significant respiratory disease.

• Genetic identification of known variant alleles that affect the pharmacokinetics or pharmacodynamics of opioid agents can enable anesthesia providers to better select the appropriate opioid and dosing regimen for an individual patient, instead of empirical selection and dosing escalation. Currently, personalized opioid selection and dosing for perioperative pain management is still a long way off. More studies are needed to improve genotype-based personalized perioperative care in children.