A dose-response curve plots a drug's effect against its (log) dose: its left-right position shows potency (EC50; lower EC50 = more potent) and its height shows efficacy (Emax, the largest effect the drug can produce). A competitive antagonist shifts the agonist curve to the right in parallel and leaves Emax unchanged; a noncompetitive or irreversible antagonist lowers Emax; a partial agonist plateaus below a full agonist's Emax and can blunt a full agonist. Graded curves show how much one patient or tissue responds; quantal curves show what percentage of a population responds, which gives the ED50, the TD50 and the therapeutic index (TI = TD50/ED50).
Tip 1: Read the y-axis first (graded vs quantal)
| Graded curve | Quantal curve | |
|---|---|---|
| Question it answers | How much effect? | How many subjects respond? |
| Y-axis | Size of the response in one patient, tissue or cell (fall in BP, % of maximal contraction) | Cumulative % of a population that reaches an all-or-none endpoint (seizure stopped, asleep, toxic effect, death) |
| Key values | EC50 (concentration giving 50% of that drug's own Emax) and Emax | ED50, TD50, LD50 (dose at which 50% of the group shows the effect, the toxicity or death) |
| What the slope tells you | How fast the effect rises with dose | How much individuals vary in sensitivity |
Why a quantal curve is S-shaped: individual sensitivities form a bell-shaped distribution, and plotting them cumulatively gives the sigmoid.
Trap: a quantal ED50 is not the same thing as a graded EC50. The ED50 is "the dose that works in half the subjects"; the EC50 is "the concentration that produces half of the maximal effect."
Tip 2: Left-right is potency, up-down is efficacy
- Affinity is how likely the drug is to occupy the receptor; intrinsic activity is how strongly the bound drug activates it. Both are set by the drug's chemical structure.
- Potency is the curve's position on the dose axis, summarized by the EC50. Further left means a lower EC50 and a more potent drug.
- Maximal efficacy (Emax) is the greatest attainable response, the drug's "ceiling effect."
- Slope is the change in response per unit of dose. A steep curve means a small dose change causes a big change in effect.
Merck's classic three-drug comparison shows why potency and efficacy are separate properties:
| Drug | Position on the dose axis | Maximal response | Reading |
|---|---|---|---|
| X | Furthest left | Full | Most potent; same efficacy as Z |
| Y | Between X and Z | Lower | More potent than Z, but less efficacious |
| Z | Furthest right | Full | Least potent; same efficacy as X |
Potency only decides the size of the dose; efficacy decides how much the drug can achieve at all, so drug Y never matches X or Z however much you give. A partial agonist can likewise be more potent than a full agonist and still be less efficacious.
| What you see | What changed | What it means |
|---|---|---|
| Curve shifts left | EC50 ↓ | More potent: less drug for the same effect |
| Curve shifts right | EC50 ↑ | Less potent: more drug for the same effect |
| Plateau drops | Emax ↓ | Less efficacious: extra drug cannot recover the maximum |
Spare receptors (receptor reserve)
A tissue has spare receptors when a full agonist can produce the maximal response while occupying only a fraction of the receptors, because the downstream signal is amplified. Three consequences are testable:
- EC50 is lower than Kd. The agonist looks more potent in the tissue than its binding affinity predicts; if the EC50 is less than the Kd, spare receptors are said to exist.
- An irreversible antagonist acts in two stages. Increasing doses first shift the agonist curve progressively to the right with the maximum preserved. Once too few functional receptors remain to reach the original maximum, the curve stops shifting and Emax falls.
- A partial agonist can look like a full agonist. With enough amplification, a partial agonist can still produce the full maximal response, so a drug that acts as a partial agonist in one tissue may act as a full agonist in another.
Tip 3: Name the curve modifier
Agonists: full, partial and inverse
| Ligand | Intrinsic activity | Given alone | Given with a full agonist | Examples |
|---|---|---|---|---|
| Full agonist | Maximal | Reaches the system's Emax | (reference curve) | Morphine, fentanyl (mu opioid) |
| Partial agonist | Submaximal | Plateaus below a full agonist's Emax, even when every receptor is occupied | Competes for the same receptors; against high full-agonist levels it pulls the response down toward its own ceiling | Buprenorphine (partial mu agonist), pentazocine |
| Inverse agonist | Negative | Stabilizes the inactive receptor, lowering constitutive (agonist-independent) activity below baseline | Blocks the agonist, much like a competitive antagonist | Pimavanserin (inverse agonist and antagonist at 5-HT2A) |
| Antagonist | None | No effect on its own | Blocks the agonist (see the next table) | Naloxone, flumazenil |
Buprenorphine binds mu receptors with high affinity and dissociates slowly but activates them less than full agonists: its respiratory depression has a ceiling, it can precipitate withdrawal if given too soon after a full agonist, and its overdose can be hard to reverse with naloxone (see the opioid use disorder guide).
Antagonists: how each one moves the agonist curve
| Antagonist type | How it binds | Agonist EC50 | Agonist Emax | Overcome by more agonist? | Example |
|---|---|---|---|---|---|
| Competitive, reversible | Same site as the agonist; binds reversibly and readily dissociates | ↑ (parallel right shift) | Unchanged | Yes (surmountable) | Naloxone vs morphine; flumazenil vs benzodiazepines |
| Irreversible | Same site, stable covalent bond (e.g., alkylation) | Unchanged, or ↑ at first if spare receptors exist | ↓ | No: function returns only as receptors are replaced | Phenoxybenzamine at alpha receptors |
| Noncompetitive | Agonist can still bind, but the antagonist reduces or prevents its action | Usually unchanged | ↓ | No (insurmountable) | Ketamine at the NMDA receptor |
- Competitive: the exam picture is curves sliding right with the same plateau; a high enough morphine dose overcomes naloxone. Flumazenil competitively blocks the benzodiazepine site on the GABA-A receptor; it does not reverse ethanol, barbiturates, general anesthetics or opioids, and it can trigger seizures in patients on long-term benzodiazepines (more in the benzodiazepine tips).
- Irreversible: phenoxybenzamine combines irreversibly with alpha-adrenergic receptors; effects of an IV dose persist for at least 3-4 days. It controls hypertension and sweating in pheochromocytoma.
- Mixed picture: a curve that shifts right and loses height suggests an insurmountable antagonist in a tissue with some receptor reserve. For USMLE, a falling Emax means a noncompetitive mechanism (or a partial agonist).
Reading the graph in 10 seconds
- Plateau the same, curve slid right? Competitive antagonist (potency ↓, efficacy unchanged).
- Plateau lower, midpoint about the same? Noncompetitive or irreversible antagonist (efficacy ↓).
- Drug alone gives a submaximal plateau, and adding it to a full agonist pulls the response down toward that plateau? Partial agonist.
- Drug alone pushes activity below baseline? Inverse agonist.
Memory hook for the receptor types: the agonist vs antagonist vs partial agonist mnemonic.
From the curves to the therapeutic index
Safety comes from quantal curves for the desired effect and the toxic effect, plotted on the same dose axis:
- Therapeutic index (TI) = TD50 / ED50 (LD50/ED50 in animal studies). The bigger the gap between the two curves, the larger the TI and the safer the drug. Merck's clinical version is the ratio of the minimum toxic concentration to the median effective concentration.
- Margin of safety = TD01 / ED99 (ToxTutor): the dose toxic to 1% of the population divided by the dose effective in 99%. It exposes overlapping curves that a 50% ratio can hide.
- Therapeutic window: the concentration range that is effective without being toxic.
- Narrow TI drugs such as warfarin, lithium and theophylline need close monitoring and careful dose adjustment because their margin of safety is small; a drug with a very high TI, such as penicillin, is generally safe.
The formula, the narrow-TI drugs and the exam traps are covered in the therapeutic index guide.
How USMLE tests it
- "More potent" does not mean "more effective." A left-shifted curve with a lower plateau belongs to a more potent but less efficacious drug.
- Two ways to lower Emax. A noncompetitive antagonist does nothing by itself; a partial agonist produces some effect by itself. Check what the drug does when given alone.
- "What restores the maximal response?" More agonist works only against a competitive antagonist. After an irreversible antagonist, recovery means waiting for new receptors.
- Precipitated withdrawal after buprenorphine or naloxone means a partial agonist or antagonist displaced a full agonist.
- "% of patients" on the y-axis means a quantal question: ED50, TD50 or TI, not EC50 or Emax.
Practice question
A 34-year-old man with opioid use disorder comes to clinic to start treatment. He last injected heroin 2 hours ago and currently has no withdrawal symptoms. He is given sublingual buprenorphine. Within 30 minutes he develops yawning, rhinorrhea, lacrimation, piloerection, abdominal cramps and diarrhea. Which property of buprenorphine best explains these findings?
A. It is a competitive mu-receptor antagonist with no intrinsic activity B. It binds mu receptors with high affinity but has lower intrinsic activity than the full agonist already bound C. It is an inverse agonist that suppresses constitutive mu-receptor signaling D. It binds an allosteric site and lowers the maximal opioid response noncompetitively E. It has greater efficacy than heroin and rapidly desensitizes mu receptors
Answer: B. This is precipitated withdrawal. Buprenorphine is a partial mu agonist with high affinity and slow dissociation, so it displaces the full agonist but activates the receptors less, and the net opioid effect falls abruptly. On the curve, adding a partial agonist to a high level of full agonist pulls the response down toward the partial agonist's lower ceiling. This is why buprenorphine is usually started only once the patient is in mild-to-moderate withdrawal, about 12 hours after heroin or another short-acting opioid and 24 hours or more after a long-acting one.
- A: that describes naloxone. Buprenorphine has real agonist activity: it treats withdrawal when started at the right time, and its respiratory depression has a ceiling.
- C: buprenorphine is a partial agonist (and a weak kappa antagonist), not an inverse agonist.
- D: buprenorphine competes with the full agonist for the same mu receptors; it is not an allosteric noncompetitive antagonist.
- E: buprenorphine has lower, not higher, efficacy; more efficacy would deepen opioid effects, not cause withdrawal.
Key takeaways
- Left-right = potency (EC50); up-down = efficacy (Emax); the two are independent.
- Graded curves give EC50 and Emax; quantal curves give ED50, TD50 and LD50 for a population.
- Competitive antagonist: parallel right shift, same Emax, surmountable (naloxone, flumazenil).
- Noncompetitive or irreversible antagonist: Emax falls and more agonist cannot fix it (ketamine at NMDA, phenoxybenzamine at alpha receptors).
- Partial agonist: submaximal Emax even at full occupancy; acts as an antagonist next to a full agonist (buprenorphine, precipitated withdrawal).
- Inverse agonist: lowers constitutive receptor activity below baseline (pimavanserin at 5-HT2A).
- Spare receptors: EC50 < Kd; irreversible blockade shifts the curve right before Emax drops.
- TI = TD50/ED50; the margin of safety (TD01/ED99) catches overlapping curves.
Sources
- Dose-Response Relationships — Merck Manual Professional Edition (2026)
- Drug-Receptor Interactions — Merck Manual Professional Edition (2025)
- ED50 — StatPearls, NCBI Bookshelf (2024)
- Pharmacological Characterization of GPCR Agonists, Antagonists, Allosteric Modulators and Biased Ligands from HTS Hits to Lead Optimization — Assay Guidance Manual, NCBI Bookshelf (2019)
- Quantification of signal amplification for receptors: the Kd/EC50 ratio of full agonists as a gain parameter — Frontiers in Pharmacology (2025)
- Phenoxybenzamine: mechanism of action — PubChem, National Library of Medicine
- Phenoxybenzamine Hydrochloride Capsules: prescribing information — DailyMed, NLM (2026)
- Flumazenil Injection: prescribing information — DailyMed, NLM (2025)
- Buprenorphine — StatPearls, NCBI Bookshelf (2024)
- Nuplazid (pimavanserin): prescribing information — DailyMed, NLM (2026)
- Ketalar (ketamine hydrochloride) Injection: prescribing information — DailyMed, NLM (2026)
- Determining the Safety of a Drug — ToxTutor, National Library of Medicine
Frequently Asked Questions
What is the difference between potency and efficacy on a dose-response curve?
Potency is where the curve sits on the dose axis and is summarized by the EC50, the concentration that produces half of the drug's maximal effect; a curve further to the left belongs to a more potent drug. Efficacy is the height of the curve, the maximal effect (Emax) the drug can produce. The two are independent: a drug can be more potent yet less efficacious, so a more potent drug only needs a smaller dose and is not necessarily better.
What does a competitive antagonist do to a dose-response curve?
A reversible competitive antagonist binds the same site as the agonist, so it shifts the agonist's curve to the right in parallel. The EC50 rises, meaning lower apparent potency, but Emax is unchanged because a high enough agonist concentration outcompetes the antagonist. That is why this block is called surmountable. Classic examples are naloxone against morphine at mu opioid receptors and flumazenil against benzodiazepines at the GABA-A receptor.
How does a noncompetitive or irreversible antagonist change the curve?
It lowers the agonist's maximal response (Emax), and adding more agonist cannot restore it. A noncompetitive antagonist lets the agonist bind but reduces or prevents its action, while an irreversible antagonist such as phenoxybenzamine bonds permanently to alpha receptors, so function returns only as receptors are replaced. In tissues with spare receptors, low doses of an irreversible antagonist first shift the curve to the right, and Emax falls once the reserve is used up.
What is a partial agonist and why can it act like an antagonist?
A partial agonist binds the same receptor as a full agonist but activates it less, so its maximal effect stays below the full agonist's even when every receptor is occupied. When a full agonist is already present, the partial agonist takes over receptors and lowers the total response toward its own ceiling. Buprenorphine, a high-affinity partial mu agonist, shows both sides: a ceiling on respiratory depression, and precipitated withdrawal if it is given too soon after a full opioid agonist.
What is the difference between graded and quantal dose-response curves?
A graded curve measures how much response one patient, tissue or cell produces as the dose rises, and it gives the EC50 and Emax. A quantal curve measures what percentage of a population reaches an all-or-none endpoint at each dose, such as seizure control, a toxic effect or death. Quantal curves give the ED50, TD50 and LD50, which are used to calculate the therapeutic index, TD50 divided by ED50.
What are spare receptors in pharmacology?
Spare receptors, also called receptor reserve, exist when a full agonist can produce the maximal response while occupying only a fraction of the receptors, because the signal is amplified downstream. As a result, the agonist's EC50 is lower than its dissociation constant (Kd). In such tissues a partial agonist can look like a full agonist, and an irreversible antagonist shifts the curve to the right before it lowers the maximal response.