Drug Interactions: Classification and Mechanisms

Overview

• Drug interactions are ubiquitous in the ICU.
• The critically ill patient receives an average of 10-20 concurrent medications, and polypharmacy is the rule rather than the exception.
• Understanding the mechanisms of drug interactions allows the intensivist to anticipate harm, exploit beneficial synergies, and rationalise complex drug regimens.
• Interactions are classified into three fundamental categories: pharmaceutical, pharmacokinetic, and pharmacodynamic.

Classification of Drug Interactions

• A drug interaction is considered clinically significant only when it results in either unexpected toxicity or a meaningful loss of efficacy.
• In a closely monitored environment such as the ICU or operating theatre, a change of at least 50% in a key pharmacokinetic or pharmacodynamic variable is generally required before a drug interaction becomes practically important.

1. Pharmaceutical Interactions

Pharmaceutical interactions are chemical or physical in nature and occur before a drug even reaches the patient, in the syringe, infusion bag, or IV tubing.

Mechanisms

• Acid-base incompatibility: Thiopental (highly alkaline, pH ~10-11) precipitates when co-administered IV with acidic solutions such as non-depolarising muscle relaxants. This is a classic example of a pharmaceutical interaction causing physical precipitation.
• Adsorption: Drugs may adsorb to plastics in IV tubing (e.g. insulin, nitroglycerin, amiodarone).
• Oxidation/reduction: Light-sensitive drugs (e.g. sodium nitroprusside) may be inactivated by ambient light exposure.

ICU Relevance

• In the ICU, multiple drugs are often co-infused through the same lumen.
• Y-site compatibility charts must be consulted.
• Incompatibility between vasopressors and alkaline solutions (e.g. sodium bicarbonate) can result in loss of active drug or particulate formation.

2. Pharmacokinetic Interactions

Pharmacokinetic interactions occur when one drug alters the absorption, distribution, metabolism, or excretion (ADME) of another, changing its plasma concentration, and therefore its effect, without directly altering receptor sensitivity.

2a. Absorption

Altered GI absorption is less relevant in the acute ICU setting (most drugs delivered IV), but clinically relevant mechanisms include:

• Binding in gut: Cholestyramine binds warfarin in the intestinal lumen, reducing bioavailability.
• Altered gastric pH: H2 antagonists and proton pump inhibitors can alter absorption of pH-dependent drugs.

2b. Distribution, Protein Binding Displacement

• Many ICU drugs are highly protein-bound (e.g. warfarin ~99% bound to albumin).
• Displacement of a drug from albumin increases the free fraction, potentially increasing effect or accelerating elimination.
• Phenylbutazone displaces S-warfarin (the more potent enantiomer) from albumin, increasing free warfarin and prolonging the INR.
• In critical illness, hypoalbuminaemia is common, meaning baseline free fractions of protein-bound drugs may already be elevated.

2c. Metabolism, Cytochrome P450 System

• The cytochrome P450 (CYP) enzyme system is responsible for oxidative metabolism of the majority of ICU drugs.
• CYP3A4 and CYP2D6 are the isoforms most relevant to opioid metabolism.

Enzyme Inhibition

Inhibition of CYP enzymes reduces metabolism of the substrate drug, causing elevated plasma concentrations and increased/prolonged effect.

Enzyme Induction

Inducers upregulate CYP enzyme expression (requires days to weeks), increasing metabolism and reducing plasma concentration of substrate drugs.

Hepatic Blood Flow

• Drugs with high hepatic extraction ratios are sensitive to changes in hepatic blood flow. Sufentanil has a higher hepatic extraction ratio than alfentanil and is therefore more affected by reductions in hepatic blood flow (as may occur in septic shock, right heart failure, or from cardiodepressant drugs).
• Cimetidine can prolong opioid effects both by CYP inhibition and by reducing hepatic blood flow.

2d. Excretion

• Altered renal filtration and tubular secretion can affect elimination of renally cleared drugs.
• In critical illness, AKI is common and dramatically alters pharmacokinetics, this is a patient-state interaction that must be recognised when prescribing renally cleared antibiotics, NMBAs, and digoxin.
• P-glycoprotein inhibitors (amiodarone, carvedilol, ciclosporin) can reduce renal tubular secretion of drugs such as naldemedine, raising plasma concentrations.

3. Pharmacodynamic Interactions

Pharmacodynamic interactions occur when one drug alters the pharmacological effect of another at the level of the receptor or effector system, without necessarily changing plasma concentrations. These are subdivided into:

• Additive
• Synergistic (supra-additive)
• Antagonistic (infra-additive)

Additionally, there is a subcategory of potentiation, where one drug with no independent action on the endpoint enhances the effect of another.

3a. Formal Definitions

$$E_{AB} = E_A + E_B \quad \text{(simple additivity, only valid for linear dose-response relationships)}$$

The Loewe additivity model is the preferred reference for "no interaction," defined as:

$$\frac{d_A}{D_A} + \frac{d_B}{D_B} = 1$$

• Where $D_A$ and $D_B$ are the doses of drugs A and B individually producing a specified effect, and $d_A$, $d_B$ are the doses in combination producing the same effect.
• This model correctly predicts that a drug combined with itself is additive, unlike the simpler effect-addition model, which fails for non-linear (sigmoidal) dose-response curves.

3b. The Isobologram

• The isobologram plots dose combinations of two drugs that produce an equal specified effect (an isobole or isoeffect curve).
• The most commonly studied isobole is the 50% isobole (ED$_{50}$), representing all combinations producing 50% of maximum effect.
• If isoboles are straight lines: interaction is additive
• If isoboles bow toward the origin: synergistic (less drug needed than expected)
• If isoboles bow away from the origin: antagonistic (more drug needed than expected)
• For clinical anaesthesia and ICU sedation, the 95% isobole is most clinically relevant, representing the concentration pairs with a 95% probability of achieving the desired endpoint (e.g. loss of responsiveness).
• Dosing to just above the 95% isobole for the desired effect, while remaining below isoboles for adverse effects (apnoea, cardiovascular depression), is the goal of multi-drug sedoanalgesia.

3c. Response Surface Models

Three-dimensional response surface models extend the isobologram concept across the full range of drug concentrations for two drugs simultaneously. The response surface plots:

• Drug A concentration (x-axis)
• Drug B concentration (y-axis)
• Probability of effect (z-axis / colour)
• This allows visualisation of all isoboles from 0-100% simultaneously, capturing complex interaction behaviour that a single isobole cannot.
• The propofol-remifentanil response surface for loss of responsiveness is one of the most well-characterised in anaesthesia.

3d. Mechanisms of Pharmacodynamic Interaction

3e. Specific Pharmacodynamic Interactions of ICU Relevance

• Inhalational agents (additive with each other): When two volatile agents are combined (e.g. sevoflurane + isoflurane), they interact in a strictly additive manner, consistent with a shared mechanism of action.
• The exception is nitrous oxide, which is infra-additive with other inhalational agents.

Intravenous agents + inhalational agents (synergistic): Propofol, benzodiazepines, and opioids interact synergistically with inhalational agents, meaning significantly less of each drug is required.

Opioids + sedative-hypnotics (synergistic):

• A subhypnotic dose of midazolam (0.07 mg/kg) reduced the ED$_{50}$ of alfentanil for anaesthetic induction from 130 μg/kg to 28 μg/kg, a >75% reduction
• A sub-analgesic dose of alfentanil (3 μg/kg) reduced the midazolam ED$_{50}$ for induction from 0.22 to 0.14 mg/kg
• A subhypnotic dose of midazolam (0.02 mg/kg) reduced the thiopental ED$_{90}$ for induction from 3.9 to 2.0 mg/kg
• The ED$_{50}$ for the midazolam-propofol combination was reduced by 37% compared to additive prediction

Ketamine, the exception (additive, not synergistic):

• Ketamine + thiopental: additive
• Ketamine + midazolam: additive
• Propofol + thiopental: additive
• These exceptions likely reflect overlapping mechanisms (both are CNS depressants but may share common molecular pathways reducing the scope for supra-additivity)

Opioids + benzodiazepines (adverse synergism, respiratory depression):

• Midazolam 0.05 mg/kg alone: no significant respiratory effect
• Fentanyl 2.0 μg/kg alone: hypoxaemia (SpO₂ <90%) without apnoea in 50% of subjects
• Combination: hypoxaemia in 11/12 subjects, apnoea in 6/12 subjects, a dramatically synergistic adverse effect

Fentanyl + diazepam (cardiovascular synergism):

• Neither drug alone caused significant haemodynamic change at the doses studied
• In combination: synergistic reduction in MAP to ~60 mmHg via reduced peripheral vascular resistance, a highly relevant interaction for ICU patients already at risk of haemodynamic instability

Warfarin, pharmacodynamic interactions:

• Synergism: Hepatic disease (reduced clotting factor synthesis), hyperthyroidism (increased clotting factor turnover), aspirin (impaired platelet function) all augment warfarin's anticoagulant effect via pharmacodynamic mechanisms
• Competitive antagonism: Vitamin K directly opposes warfarin's mechanism by restoring substrate for the vitamin K epoxide reductase cycle
• Physiological control loop: Hereditary resistance to warfarin involves mutation of vitamin K reactivation cycle molecules

ICU Relevance

Beneficial Synergism in Sedoanalgesia

• ICU sedation strategies exploit pharmacodynamic synergism.
• Combining a propofol infusion (GABA$_A$ agonist) with an opioid infusion (remifentanil, fentanyl) achieves adequate sedoanalgesia at substantially lower doses of each agent than would be required if used alone.

This reduces:

• Haemodynamic depression from propofol
• Opioid-induced ileus and immunosuppression
• Prolonged neuromuscular recovery

Dangerous Synergism, Respiratory Depression

• Opioid-benzodiazepine co-prescription outside the mechanically ventilated patient must be approached with extreme caution.
• In spontaneously breathing ICU patients (e.g. post-extubation), the synergistic respiratory depressant interaction between opioids and benzodiazepines can precipitate life-threatening apnoea at doses that individually appear safe.

Pharmacokinetic Interactions in Critical Illness

Clinically Significant Threshold

• A drug interaction is clinically significant in the ICU setting when it produces ≥50% change in a key pharmacokinetic or pharmacodynamic variable.
• Minor interactions are common but rarely require action in a well-monitored environment with titrated infusion therapy.

Pharmaceutical Incompatibilities in ICU Practice

Summary Framework for the Exam

When asked to classify drug interactions, always structure your answer using the three-category framework:

1. Pharmaceutical: chemical/physical incompatibility
2. Pharmacokinetic: altered ADME (absorption, distribution, metabolism via CYP induction/inhibition, excretion)
3. Pharmacodynamic: altered receptor-level effect (additive, synergistic, antagonistic, potentiation)

• For pharmacodynamic interactions, be able to define Loewe additivity, describe the isobologram, and give specific numerical examples of synergism from the propofol-opioid and midazolam-opioid literature.
• Understand that in the ICU, both beneficial synergism (sedoanalgesia) and harmful synergism (respiratory depression, cardiovascular collapse) are the direct clinical consequences of these mechanisms.