Discussion
The girl remained asleep on propofol, and showed no observable response to the intravenous opioids administered. As this was an elective procedure, one might question whether the intervention ought not to have been aborted, given the complete absence of an opioid response. However, experience indicates that postoperative pain after the procedure in question can be managed successfully without opioids. With the routine use of long-acting local anaesthetic and with supplementary paracetamol and NSAIDs, pain is rarely an issue. Should any problems have arisen, preparations were in place for the patient to remain in the observation unit for the first 24 hours. On this basis, it was decided to proceed with the operation.
It soon became clear that the patient lacked a normal opioid response. The pure mu (μ) agonists remifentanil and fentanyl had no effect, even in large doses. It was suspected even at this stage that there might be abnormalities at the receptor level. Morphine also acts to some degree via the kappa (κ) and delta (δ) receptors, and it is possible that higher doses of morphine might have had an effect. The analgesic effects of opioids are mediated mainly via μ-opioid receptors, and μ-receptor-mediated intracellular signalling is G-protein coupled. Previously, μ-opioid receptors were subdivided into μ1 and μ2 receptors, and it was thought that opioid analgesia was mediated by μ1 stimulation, whereas many of the adverse effects were the result of μ2 stimulation. This has subsequently been refuted. The general consensus now is that there is only one type of μ-opioid receptor, and that subtyping of opioid receptors should be avoided (4, 5) .
The gene for the μ-opioid receptor, OPRM1 , exhibits significant polymorphism. This is thought to partly explain the large variation in the clinical effects of opioids (6) . Individual differences in uptake, metabolism and transport across the blood-brain barrier also contribute substantially to between-patient differences in clinical opioid response (7) . In our patient, on the other hand, it was an inactivating mutation in the OPRM1 gene that was responsible. The mutation in question had been described previously by Skorpen et al . in Trondheim (2) , and leads to a change in the amino acid at position 181 from arginine to cysteine (2) . This amino acid is located in the intracellular domain of the receptor, and the mutation prevents transduction of extracellular stimulation into an intracellular response (Figure 1). Homozygous carriers of this mutation, such as the girl in this case report, will therefore not respond clinically to the opioids that are most commonly used in anaesthesia, such as fentanyl, alfentanil and remifentanil. Other opioids will also have no effect on pain, unless they additionally stimulate kappa and delta receptors.
Figure 1 An opioid receptor with an Arg181Cys mutation. Despite proper binding of opioids (O) to the μ-opioid receptor, intracellular signal transduction is not triggered. The receptor is inactive ('signalling dead') (2 ).
The OPRM1 mutation that had been homozygously inherited by the girl in this case report, has an allele frequency of 0.25 % in the European population (8) and approximately 1/200 (0.5 %) in the Norwegian population, see figures published by Skorpen et al . (2) . This means that the frequency of homozygous carriers of the Arg181Cys mutation in OPRM1 is about 1 in 40 000 (2) . Being completely unresponsive to opioids as a result of this mutation is therefore rare, but of critical significance. Skorpen et al . also demonstrated a decreased opioid response in heterozygous carriers of the Arg181Cys mutation. Heterozygous carriers comprise about 1 % of the population, and will therefore regularly be encountered by clinicians in the context of pain management.
A pharmacogenetic panel for investigation of the opioid response, which incorporates both OPRM1 (including the Arg181Cys mutation) and relevant CYP enzymes involved in opioid metabolism, is now available as a routine analysis at the Department of Psychopharmacology, Diakonhjemmet Hospital. To what extent patients should be screened for the mutation in clinical practice is an issue for discussion in the field. As a basic principle, it would be sensible to consider pharmacogenetic testing of patients with a very unusual opioid response following discussion with specialists in pain management. An important point in this context is that patients with long-term use of opioids for chronic pain, such as cancer patients, may develop strong opioid tolerance that is unrelated to mutation of the OPRM1 gene. Skorpen et al . have shown that an increased frequency of the Arg181Cys mutation may be anticipated in patient groups with a poor opioid response. Further research is required to determine how often pharmacogenetic variation may account for an inadequate opioid response.
The pharmacogenetic analysis of OPRM1 explained the absence of a therapeutic response to opioids in our patient. The results of the genetic testing are also important for the future: the girl in this case report will encounter challenges should she require acute pain relief later in life. The use of opioids is such an integral part of the acute treatment of moderate to severe pain that patients may be subjected to a trial and error approach to achieve pain relief, unless clinicians are aware that they fully lack an opioid response. One of the aims of the new Norwegian 'summary care record' is for this type of critical information to be available irrespective of which hospital in Norway the patient is admitted to. However, this requires doctors to manually enter information into the system. As of yet, there is no functionality to link information from the summary care record (e.g. genotype) to prescriptions, in a way that would enable automatic alerts to be generated to assist with clinical decision-making.
Patients who are homozygous for the Arg181Cys mutation in OPRM1 will require personalised pain management. Prior to any surgery, the surgeon and anaesthesiologist should agree on a strategy for perioperative management and post-operative pain relief. Infiltration anaesthesia and peripheral or central blocks will be helpful where these are possible. A plan should be in place for managing breakthrough pain and any failure of a blockade. In addition to basic analgesia with paracetamol and/or NSAIDs, adjuvant agents such as corticosteroids and clonidine may be indicated. For stronger pain, ketamine in combination with a benzodiazepine or small doses of propofol would be a natural choice. In hospitals, the use of intravenous ketamine is often limited in practice to specialist wards. For acute painful conditions that would normally require opioids, it is probably reasonable for these patients to initially be placed under observation with close follow-up by an anaesthesiologist or others with special expertise in pain management.