Pharmacology of opioids

This chapter tries to address Section K4(ii) of the 2017 CICM Primary Syllabus, which expects the exam candidate to "describe the pharmacology of opiates as a class". There's also a bit of Section K3(iii) here, where we "describe the physiology of opioid and NMDA receptors". NMDA receptors are covered well enough in the Ode de Ketamine, so this chapter deals with opiates exclusively. A vast number of questions in the past exam papers have asked for either a comparison of two opiates, or a discussion of their antagonists, or for some detail about them as a class:

  • Question 12 from the first paper of 2021 (oxycodone)
  • Question 17 from the first paper of 2020 (fentanyl vs. morphine infusion)
  • Question 13 from the second paper of 2017 (fentanyl vs morphine)
  • Question 12 from the first paper of 2017 (oxycodone)
  • Question 12 from the second paper of 2016 (fentanyl vs morphine)
  • Question 24 from the first paper of 2014 (mechanism of opioid effect)
  • Question 15 from the first paper of 2013 (opioid antagonists)
  • Question 16 from the first paper of 2011 (morphine, fentanyl and remifentanil)
  • Question 12(p.2) from the first paper of 2010 (mechanism of opioid effect)
  • Question 12 from the first paper of 2009 (pharmacology of spinal opioids)

In summary:

  • Opiates are drugs derived directly from opium (e.g. morphine, codeine, and heroin);
  • Opioids are a broad class of opiate analogues that have morphine-like activity
  • Absorption of opioids orally is generally excellent
  • First pass effect tends to decrease the bioavailability of most of them
  • Their metabolism is mainly hepatic, and many of them have active metabolites (morphine, codeine, buprenorphine, heroin)
  • Opioid receptors are G-protein coupled receptors, mainly situated on the presynaptic membrane, which increase potassium conductance and decrease calcium conductance. The net effect of their activation is to hyperpolarise the membrane and prevent neurotransmitter release.
  • Their mechanism of analgesic action is mainly related to the inhibition of glutamate release from  primary pain afferent neurons in the spinal cord
  • A secondary effect is the potentiation of the descending inhibitory pathways which regulate pain (which originate in the midbrain)

It's another one of those situations where one really does not need to make specific recommendations regarding published peer-reviewed material because the Google Scholar search for "opioids" produces a dozen articles titled "Opioids". If a reader needed a curated selection of literature, a good starting point would be Inturrisi (2002) for "classical" opioids and Armenian et al (2018) for fentanyl and novel synthetic analogs.  For the reader with institutional access to Springer-Verlag, the best recommendation would have to be Crow et al (2021), a chapter from Pasqual & Gaulton's  Opioid use in Critical Care (2021), or the chapter on opioids by Zöllner & Stein from the Handbook of Experimental Pharmacology (2006). Both of them combine the thoroughness of a textbook with the fastidious referencing of a scholarly thesis, but be warned: that bibliography section is a powerful rabbit hole generator. 

Opium, opiates and opioids

Opium, from ὄπιον  ("milk of the poppy"), is not a chemical term but a colloquial one, to describe a "spontaneously coagulated latex which exudes when the partly ripened capsules of Papaver somniferum L. are lanced on the living plant" (Dunnicliff, 1937). This raw natural plant ingredient seems to have disappeared from the chemist's shelves at some point in the early 20th century. Its main ingredient, in case anyone is interested, was morphine - it was 10-15% morphine by weight, as well as, codeine (1%–3%), noscapine (4%–8%), papaverine (1%–3%), and thebaine (1%–2%). The morphine Hauptbestandteil was isolated and named by Friedrich Wilhelm Sertürner who consumed some uncontrolled quantity of it in 1817. "When he awoke hours later, he realised that this compound was safe for human consumption", knowingly nodded Krishnamurti & Rao (2016). 

In fact, though raw opium in its original form was no longer available, its presence in pharmaceuticals remained very popular, and it remained on the list of ingredients in many off-the-shelf products in pharmacies. The graph below (from Anderson & Berridge, 2000) is remarkable mainly by the tail of it, where even by 1998 the number of "official" preparations containing opium was still not zero.

number of official preparations containing opium, from Anderson & Berridge (2000)

Thus, pharmacists and doctors all seem to have had some unhealthy attachment (call it an addiction?) to this substance, which to some extent explains the popularity of the name root in etymology. It does also help that the word "opiates" itself has some strong bohemian energy, whereas "natural phenanthrene and benzylisoquinoline alkaloids" does not. Thus, official terminology holds that:

"Opiates are drugs derived directly from opium (e.g. morphine, codeine, and heroin); while opioids are a broad class of opiate analogues that have morphine-like activity (e.g. methadone)."

This seems like an attempt to classify these substances by their origin; as most morphine manufactured worldwide is still made from poppy plants (though they are no longer delicately lanced and milked - in this cyberpunk dystopia of the 21st century, some kind of brutal machine mechanically separates them and turns them into a fibrous pulp from which the raw drug is extracted by the ton). This brings us to:

Classification of opiates and opioids

Like with anything complex and subtle in medicine, there is a handful of ways of doing the same thing, and none of them is considered definitive or correct. Thus, there are multiple ways to classify opioid receptor agonists. The best overview of these ways can be found in  Crow et al (2021), from where the following has been stolen:

Traditional classification Chemical (structural) classification Functional classification

Natural compounds

  • Codeine
  • Morphine
  • Papaverine
  • Thebaine

Semi-synthetic compounds

  • Buprenorphine
  • Hydrocodone
  • Hydromorphone
  • Oxycodone
  • Oxymorphone

Synthetic compounds

  • Alfentanil
  • Butorphanol
  • Fentanyl
  • Levorphanol
  • Meperidine
  • Methadone
  • Methylnaltrexone
  • Nalbuphine
  • Naltrexone
  • Propoxyphene
  • Remifentanil
  • Sufentanil
  • Tramadol

Natural phenanthrenes

  • Morphine
  • Codeine

Semi-synthetic phenanthrenes

  • Buprenorphine
  • Hydrocodone
  • Hydromorphone
  • Oxycodone
  • Oxymorphone

Synthetic phenanthrenes

  • Butorphanol
  • Levorphanol
  • Methylnaltrexone
  • Nalbuphine
  • Naloxone
  • Naltrexone

Diphenylheptanes

  • Methadone
  • Propoxyphene

Piperidines

  • Meperidine
  • Remifentanil
  • Fentanyl
  • Sufentanil
  • Alfentanil

Phenylpropylamines

  • Tramadol
  • Tapentadol

Agonists

  • Alfentanil
  • Codeine Fentanyl
  • Hydrocodone
  • Hydromorphone
  • Meperidine
  • Methadone
  • Morphine
  • Oxycodone
  • Oxymorphone
  • Propoxyphene
  • Remifentanil
  • Sufentanil

Partial agonists

  • Buprenorphine
  • Butorphanol
  • Nalbuphine
  • Pentazocine
  • Tramadol

Centrally acting antagonists

  • Naloxone
  • Naltrexone

Peripherally acting antagonists

  • Alvimopan
  • Methylnaltrexone
  • Naldemedine
  • Naloxegol

Pharmacokinetics of opioids as a class

Absorption and bioavailability

For the most, opioids are well absorbed from an enteric form, and undergo significant first-pass metabolism to yield low-ish bioavailability figures.

Name Routes of administration Absorption Bioavailability
Morphine Oral, IV, epidural, intrathecal, transdermal, subcutaneous, IM Well absorbed orally 30% bioavailability
Codeine Oral, IV Well absorbed orally 50% bioavailability
Oxycodone Oral, IV Well absorbed orally 80% bioavailability
Heroin Oral, IV Well absorbed orally Dose-dependent oral bioavailability from 40% with low doses to 65% with high doses
Buprenorphine Oral, IV, topical Well absorbed orally 15% bioavailability due to high first pass effect
Hydromorphone Oral, IV Well absorbed orally 50% oral bioavailability
Alfentanil IV Well absorbed orally 50% oral bioavailability
Methadone Oral, ...IV?... Well absorbed orally 80% oral bioavailability, but with a wide range of interindividual variation
Tapentadol Oral, IV Well absorbed orally 20% oral bioavailability
Tramadol Oral, IV 100% absorbed orally 70% oral bioavailability, but increases to 90-100% with sustained dosing because of hepatic enzyme saturation
Fentanyl Oral, IV, epidural, intrathecal, transdermal, subcutaneous, IM Well absorbed orally;  mucosal absorption is poor. Transdermal absorption is slow. Oral bioavailability is 33%.
Remifentanil IV, intranasal Oral absorption is poor. Mucosal absorption is relatively rapid and it can be used intranasally Oral bioavailability is poor- thought to be near 0%.

Of these agents, the standouts are tapentadol and buprenorphine (with very high first pass metabolism taking out 80-90% of the active drug before it gets to the circulation) and remifentanil which does not survive for long in the circulation.

Lipid solubility of opioids

Opioids are - for the majority of them - weak bases with a pKa close to 8. This means that at physiological pH many of them will have a substantial fraction of the drug present in a lipid-soluble form.

Name pKa Ionisation and lipid solubility
Morphine pKa 8.0 23% is unionised at pH 7.4; octanol-water partition coefficient ~ 1.42
Codeine pKa 8.21 relatively hydrophilic
Oxycodone pKa 8.5 quite hydrophilic
Heroin pKa 7.95 highly lipophilic (more than morphine)
Buprenorphine pKa 8..5 highly lipophilic
Hydromorphone pKa 8.6 octanol-water partition coefficient ~ 1.28
Alfentanil pKa 6.5 90% unionised at pH 7.4; octanol-water partition coefficient ~ 128
Methadone pKa 9.2 1% unionised at pH 7.4, octanol-water partition coefficient ~ 117
Tapentadol pKa 9.34 highly lipophilic
Tramadol pKa 9.41 highly lipophilic
Fentanyl pKa 8.4 9% is unionised at pH 7.4. Highly lipid soluble: octanol:water partition coefficient is 717
Remifentanil pka 7.26 42% is unionised at pH 7.4. Highly lipid soluble: octanol:water partition coefficient is 17.9

This lipid solubility metrics used in this table are somewhat variable, as not everybody reports or measures the lipid solubility of these drugs in the same way, but the main point is that most of these drugs are highly variable in their lipid solubility, and this gives rise to clinically important differences between them:

  • High lipid solubility is associated with high analgesic efficacy. This makes sense, as it would be directly related to the rate of penetration through the blood brain barrier. 
  • High lipid solubility is associated with more rapid onset of action, for the same reasons.
  • Low lipid solubility is associated with a longer duration of action, particularly from the spinal space, but also more generally (DeCastro et al, 2012). For example, morphine has a shorter half-life than fentanyl, but because of its lower lipid solubility, it takes longer to diffuse out of the CNS, and therefore has a longer duration of analgesic activity.

Distribution and protein binding of opioids

All opioids are rapidly and widely distributed, except for remifentanil which basically has no time in which to do so, as its molecules are so short-lived.

Name Volume of distribution (L/kg) Protein binding (%)
Morphine 1-6 20-35%
Codeine 3-6 7-25% 
Oxycodone 2.6 38–45% 
Heroin 5 40%
Buprenorphine 87-197 96% 
Hydromorphone 4 8-19% 
Alfentanil 0.4-1.0 88-92% 
Methadone 1-8  85-90%
Tapentadol 7 20% 
Tramadol 2.6-2.9 20% 
Fentanyl 4-6  81-94% 
Remifentanil 0.1 70% 

Metabolism and elimination of opioids

All opioids undergo hepatic metabolism except remifentanil, which is broken down by plasma esterases. The vast majority of them have no active metabolites, but notably heroin and codeine have morphine as one of their metabolic breakdown products. Most of these metabolites are renally cleared, and would be expected to accumulate in renal failure.

Name Metabolism Notable metabolites
Morphine Hepatic morphine 6-glucuronide, an active metabolite
Codeine Hepatic morphine
Oxycodone Hepatic noroxycodone and oxymorphone, both active metabolites
Heroin Hepatic morphine and 6-acetylmorphine, which is an active metabolite
Buprenorphine Hepatic norbuprenorphine, an active metabolite
Hydromorphone Hepatic inactive metabolites
Alfentanil Hepatic inactive metabolites
Methadone Hepatic inactive metabolite EDDP
Tapentadol Hepatic inactive metabolites
Tramadol Hepatic O-desmethyltramadol, an active metabolite
Fentanyl Hepatic inactive metabolites
Remifentanil Plasma esterases inactive metabolites

Notable exceptions to the above include methadone, for which at high doses renal excretion becomes an important pathway of clearance. Apparently this is seen in daily doses exceeding 55mg.

Pharmacokinetics of opioid infusions

Question 17 from the first paper of 2020 asked for "the advantages and disadvantages of the use of an intravenous infusion of fentanyl in comparison to morphine", with the examiners complaining that the candidates did not approach the question with sufficient clinical maturity. To be fair, these are people who were interns a couple of years ago, but that does bring up an important matter which has real-life relevance for intensive care people. In other words, you should be able to answer this question as a junior ICU trainee. The importance of this question is escalated considerably by the knowledge that Rinaldo Bellomo published an article on this exact topic in his journal, approximately a year before it appeared in the exam. 

  • Pharmacetical  and administrative considerations
    • The cost of morphine is approximately half the cost of fentanyl, and the longer duration of effect could make it even more cost-effective.
    • The prolonged action of its effect, however, may increase the duration of ventilation of ICU stay, increasing the overall healthcare cost.
    • Fentanyl is 100 times more potent than morphine, which means theoretically it should be more cost-effective to use it for analgesia
  • Compartment distribution and context-sensitive half time  
    • Fentanyl is widely and rapidly distributed, accumulating in tissues with sustained infusion
    • Because of tissue compartment distribution, context-sensitive half time for fentanyl is markedly prolonged following a long course of use as analgo-sedation (eg. after a 4-hour infusion, context sensitive half time is 200 minutes)
    • Morphine is distributed less widely, and its context-sensitive half-time is independent of the duration of infusion
    • However, the accumulation of morphine metabolites in critically ill patients may still produce a prolonged effect sustained long after the infusion has ceased.
  • Protein-binding
    • Fentanyl is 80-90% protein bound; with hypoalbuminaemia of critical illness, the free fraction will be increased, potentiating the clinical effects
    • Morphine is only 30-35 % protein bound and therefore less affected by hypoproteinaemia
  • The effect of lipid solubility on offset of effect:
    • Morphine has relatively poor lipid solubility as compared to fentanyl.
    • The clearance of morphine from the CNS is therefore delayed, producing a prolonged  duration of effect 
    • Fentanyl has excellent lipid solubility, and is cleared rapidly from the CNS
    • The rapid clearance of fentanyl is more likely to lead to opioid withdrawal following a long-term sustained infusion as part of ICU sedation (opioid withdrawal is an under-recognised contributor to ICU delirium)
  • Differences in metabolism
    • Morphine is metabolised into morphine6-glucuronide, an active metabolite that accumulates in renal failure.
    • Fentanyl does not have active metabolites
  • Pharmacodynamic differences
    • M​​​​​orphine may act as a direct vasodilator through its histaminergic effects, which may be beneficial (eg. in CCF) or a disadvantage (eg. septic shock), whereas fentanyl does not have direct cardiovascular effects.

Pharmacokinetics of intrathecal opioids

"The complications of spinal opioids are not an uncommon reason for admission to intensive care thus it is important candidates understand their pharmacology", the college commented on Question 12 from the first paper of 2009. Obviously, it was not that important, as the question has not been repeated ever since.  The chapter on the pharmacokinetics of drugs in the subarachnoid and epidural space would probably have more relevance to the revising trainee, as its focus is thoroughly non-opioid, focusing more on the weirdness of administering anything directly into the CSF. Bernard et al (2002) give a good opioid-specific account of what happens. 

Domain Epidural opioid pharmacokinetics: Intrathecal opioid pharmacokinetics:
Pharmaceutics

Some opioid formulations cannot be administered intrathecally or via epidural because of specific excipients, eg. remifentanil comes as a lyophilized powder with a glycine buffer, and IV morphine comes with preservatives

Absorption into target site

Rate of diffusion into CSF is slower, and depends on Fick's Law of Diffusion (volume, i.e.  surface area of available meninges, concentration, protein binding and free fraction, as well as CSF flow rate/turbulence)

Rate of diffusion into target tissue is rapid:

  • High CSF concentration
  • Short diffusion distance (2-4mm)
Local
distribution 

Epidural spaces are irregular, segmental, and the injected material encircles the dural sack.

Depends on baricity (density of injectate relative to CSF)
Systemic distribution

Two-compartment model: Rapid early distribution (into epidural fat) and then slowly back out.

Slow absorption; increased half life

Lipid solubility

The effect of the lipid solubility of opioids on their spinal and epidural effect:

  • High lipid solubility (eg. fentanyl)
    • More rapid onset of effect
    • More rapid offset of effect
    • More rapid uptake into capillaries, and therefore faster clearance from the CNS spaces
  • Lower lipid solubility (eg. morphine)
    • Slower onset of effect
    • Slower offset of effect
    • Slow clearance from the CNS, and prolonged clinical effect
      (eg. hypotension). 
Metabolism and elimination Because of the haemodynamic effects of spinal (anaesthetic) drugs, the perfusion of liver and kidneys may be decreased, which could delay clearance. 
Pharmacodynamics

Intrathecal opioids affect spinal μ-opioid receptors as well as being redistributed to the brainstem. Epidural opioids are also absorbed systemically to have an effect similar to IV administration

Side-effects
  • Hypotension
  • Pruritis
  • Urinary retention
  • Unexpectedly prolonged action
  • Respiratory depression with higher spinal levels

Physiology of the opioid receptor

Opioid receptors are a large family, and it would likely be pointless to discuss each one and their distribution, particularly as this is a rapidly developing field and material current at the time of writing would be superseded by new publications within about fifteen minutes. Thus, only timelessly generic things will be mentioned here, and the reader with thirst for more pharmacodynamic information will be referred to excellent review articles such as Waldhoer et al (2004).

Some basic facts:

  • Opioid receptors are transmembrane-spanning G-protein-coupled receptors
  • Their natural ligands are endorphins and encephalins 
  • There are four main types: μ, κ, δ and the NOP receptors (nociceptin-orphanin)
  • Activation of the receptors is inhibitory:
    • The receptors are directly coupled to potassium and voltage-sensitive calcium channels
    • By activating, they increase potassium conductance and decrease calcium conductance
    • Increased potassium conductance leads to membrane hyperpolarisation
    • Closure of calcium channels decreases the availability of intracellular calcium which is essential for neurotransmitter release
    • Ergo, the net effect of binding presynaptic opioid receptors is to decrease synaptic neurotransmission by preventing neurotransmitter release

Each receptor subtype is responsible for specific clinical effects:

  • μ-opioid receptors are responsible for much of the analgesia, but also for the respiratory depression, constipation and cardiovascular effects
  • δ-opioid receptors seem to be involved in respiratory depression, constipation and mood
  • κ-opioid receptors are implicated in the sedation and confusion seen with opioid use
  • NOP receptors seem to have an anti-analgesic, pronociceptive effect, and NOP receptor antagonists have analgesic effects

The receptors are distributed unevenly though the CNS. μ-opioid receptors, which are probably the most important clinically, are located all over the CNS, but particularly in the following key areas:

  • dorsal horn of the spinal cord: μ-receptors are present presynaptically on primary afferent neurons (where they have an inhibitory influence on neurotransmission)
  • Periaqueductal grey matter: This part of the brainstem sends descending efferents which act to inhibit nociceptive transmission in afferent fibres; μ-receptors remove some of the GABA-ergic inhibitory tone which regulates this descending inhibition

Clinical effects of opioids

Question 24 from the first paper of 2014 asked for some detailed information about the mechanism of the pharmacodynamic effect of opioids. Analgesia, respiratory depression and constipation were specifically of interest. For a brief and to-the-point overview of the former, the reader is redirected to the Australian Prescriber, where the matter is dealt with by Loris Chahl (1996). Individuals with a high pain tolerance could also be pointed to Gavril Pasternak (2005), or to the side-effects section of Zöllner & Stein, but most normal people would agree that this would be too much for a single CICM exam answer. Useful content has been scraped out of these texts and arranged below.

Mechanism of analgesia

Mechanism of opioid-induced analgesia is by two main effects:

  • Spinal μ-receptor effect:
    • Presynaptic inhibition of neurotransmitter (glutamate) release from primary nociceptive afferent neurons
    • Thus, decreased transmission of nociceptive signals to the dorsal horn neurons
  • Midbrain μ-receptor effect:
    • Inhibition of GABA-ergic input into the periaqueductal grey matter
    • Thus, decreased inhibition of descending efferent regulatory fibres which project from the periaqueductal grey matter to the dorsal horn
    • Thus, increased descending inhibition of dorsal horn neurons.

This is also the mechanism of action of endogenous endorphin substances, which also bind to the same receptors. Their physiological role is also related to pain, and - weirdly - to satiety, sexual arousal, goal-oriented incentive selection and maternal behaviour. They appear to be released by the pituitary gland in response to pain and injury, which suggests some sort of "natural analgesic" role.

Mechanism of opioid-induced respiratory depression

This is mediated by two main mechanisms: 

  • Brainstem μ-receptor effect:
  • Medullary chemoreceptor (also a μ-receptor effect):
    • Opioids decrease the sensitivity of the medullary chemoreceptors to hypoxia and hypercapnia
    • This blunts the normal response to hypoventilation

The existence of this effect is frankly baffling, because, just as with the oculocardiac reflex which can stop the heart in response to eye pressure, it is impossible to imagine what normal physiological role it might play. Why would the human organism want this to happen? 

Mechanism of opioid-induced constipation:

  • δ- and κ-opioid receptors are expressed in the enteric nervous system (Townsend et al, 2004)
  • This produces direct activation of smooth muscle cells, leading to:
    • Constant tonic contraction of smooth muscle
    • Inhibition of normal intestinal secretions
    • Inhibition of peristalsis
  • Little tolerance seems to develop to this effect

Other opioid side effects

  • Bradycardia and ablation of cardiac reflexes
  • ​​​​​​Histamine release and vasodilation, which occurs by the displacement of histamine from tissue mast cells, but appears to be a completely non-immune phenomenon. 
  • Nausea is a direct effect of opiates on the area postrema
  • Sedation is not well explained, but is definitely a recognised side effect, and appears to potentiate the sedating effects of other drugs
  • Cough suppression is mediated by the inhibitory effect of opioids on the medullary controllers responsible for the mechanism of the cough reflex. Bizarrely, this effect does not seem to be related to the magnitude of the actual opioid effects of the drugs on their receptors; for example, dextromethorphan has nil analgesic activity, but is a potent cough suppressant. 
  • Miosis seems to be a direct opioid effect on the Edinger-Westphal nucleus of the midbrain, which is observed at even subanalgesic doses, and which is also not subject to tolerance
  • Chest wall rigidity, that is really just the manifestation of whole body rigidity which is the most apparent to the chest-focused anaesthetist. This appears to be the consequence of opioids inhibiting GABA release in the striatal region of the basal ganglia, i.e a Parkinson-like effect. It only seems to occur with very high doses.

References

Zöllner, C., and C. Stein. "Opioids." Handbook of Experimental Pharmacology (2006): 31-63.

Crow, Jessica R., Stephanie L. Davis, and Andrew S. Jarrell. "Pharmacology and Pharmacokinetics of Opioids in the ICU." Opioid Use in Critical Care. Springer, Cham, 2021. 31-64.

Cata, Juan P., and Shreyas P. Bhavsar. "Pharmacology of opioids." Basic Sciences in Anesthesia. Springer, Cham, 2018. 123-137.

Armenian, Patil, et al. "Fentanyl, fentanyl analogs and novel synthetic opioids: a comprehensive review." Neuropharmacology 134 (2018): 121-132.

Inturrisi, Charles E. "Clinical pharmacology of opioids for pain." The Clinical journal of pain 18.4 (2002): S3-S13.

Dunnicliff, H. B. "Chemistry of Indian Opium." Nature (1937). 

Krishnamurti, Chandrasekhar, and SSC Chakra Rao. "The isolation of morphine by Serturner." Indian journal of anaesthesia 60.11 (2016): 861.

Sertürner, F. "Ueber das Morphium als Hauptbestandteil des Opiums." Annalen der Physik 25 (1817): 8-89.

Lueck, Anna Elizabeth Stahler. "Extraction and Purification of Morphine from Opium Poppies." (2014).

De Castro, Joris, Jacques Meynadier, and Michael Zenz. Regional opioid analgesia: physiopharmacological basis, drugs, equipment and clinical application. Vol. 20. Springer Science & Business Media, 2012.

Casamento, Andrew, and Rinaldo Bellomo. "Fentanyl versus morphine for analgo-sedation in mechanically ventilated adult ICU patients." Critical Care and Resuscitation 21.2 (2019): 76-83.

Dahlström, B. E. N. G. T. "Pharmacokinetics and pharmacodynamics of epidural and intrathecal morphine." International anesthesiology clinics 24.2 (1986): 29-42.

Bernards, Christopher M. "Understanding the physiology and pharmacology of epidural and intrathecal opioids." Best Practice & Research Clinical Anaesthesiology 16.4 (2002): 489-505.

Waldhoer, Maria, Selena E. Bartlett, and Jennifer L. Whistler. "Opioid receptors." Annual review of biochemistry 73.1 (2004): 953-990.

Pasternak, Gavril W. "Molecular biology of opioid analgesia." Journal of pain and symptom management 29.5 (2005): 2-9.

Koob, George F. "Drugs of abuse: anatomy, pharmacology and function of reward pathways." Trends in pharmacological sciences 13 (1992): 177-184.

Chahl, Loris A. "Opioids-mechanisms of action." (1996).

Manzke, Till, et al. "5-HT4 (a) receptors avert opioid-induced breathing depression without loss of analgesia." Science 301.5630 (2003): 226-229.

Townsend, DeWayne, Philip S. Portoghese, and David R. Brown. "Characterization of specific opioid binding sites in neural membranes from the myenteric plexus of porcine small intestine." Journal of Pharmacology and Experimental Therapeutics 308.1 (2004): 385-393.

Shook, Jennifer E., W. David Watkins, and Enrico M. Camporesi. "Differential roles of opioid receptors in respiration, respiratory disease, and opiate-induced respiratory depression." Am Rev Respir Dis 142.4 (1990): 895-909.